Stabilization of Myc through Heterotypic Poly-Ubiquitination by mLANA Is Critical for γ-Herpesvirus Lymphoproliferation

Host colonization by lymphotropic γ-herpesviruses depends critically on expansion of viral genomes in germinal center (GC) B-cells. Myc is essential for the formation and maintenance of GCs. Yet, the role of Myc in the pathogenesis of γ-herpesviruses is still largely unknown. In this study, Myc was shown to be essential for the lymphotropic γ-herpesvirus MuHV-4 biology as infected cells exhibited increased expression of Myc signature genes and the virus was unable to expand in Myc defficient GC B-cells. We describe a novel strategy of a viral protein activating Myc through increased protein stability resulting in increased progression through the cell cycle. This is acomplished by modulating a physiological post-translational regulatory pathway of Myc. The molecular mechanism involves Myc heterotypic poly-ubiquitination mediated via the viral E3 ubiquitin-ligase mLANA protein. EC5SmLANA modulates cellular control of Myc turnover by antagonizing SCFFbw7 mediated proteasomal degradation of Myc, mimicking SCFβ-TrCP. The findings here reported reveal that modulation of Myc is essential for γ-herpesvirus persistent infection, establishing a link between virus induced lymphoproliferation and disease.


Introduction
Myc is a transcription factor that enhances the expression of genes involved in cellular growth and proliferation. Hence, it is not surprising that viruses have evolved mechanisms to modulate Myc to promote their own life cycle. Myc heterodimerizes with Max, through a basic region/helix-loop-helix/leucine-zipper domain, to regulate the transcription of specific E-box-containing genes in response to mitogenic stimuli. Myc functions as a universal amplifier of gene expression by promoting the transcriptional elongation of RNA polymerase II driving biomass accumulation and enhanced cellular bioenergetic pathways [1,2,3]. The expression of c-myc is tightly regulated with extremely short halflives for mRNA and protein. In non-transformed cells, Myc is continuously subjected to ubiquitination and proteasomal-degradation, resulting in a highly unstable protein with a half-life of about 15-20 minutes [4]. Several mechanisms of Myc regulation have been identified that operate at the level of protein stability. The best characterized mechanism involves the interplay between phosphorylation at two specific residues and ubiquitination. Phosphorylation at serine (S) 62 by extracellular signal-regulated kinase (ERK) stabilizes Myc resulting in enhancement of its transcription activity. In contrast, phosphorylation of Myc at threonine (T) 58 by glycogen synthase kinase 3 (Gsk-3), which is dependent on previous phosphorylation of Myc at S62, leads to proteasomal degradation of Myc [5]. The mechanism involves the assembly of homotypic poly-ubiquitin chains on Myc specifically dependent on lysine (K) 48 linkage by SCF (Skp1/Cul/Fbox) Fbw7 [6,7]. Myc turnover by SCF Fbw7 is antagonized by polymerization of mixed heterotypic poly-ubiquitination chains via SCF b-TrCP on the N-terminus of Myc [8]. Thus, SCF Fbw7 and SCF b-TrCP assemble different K-linkage poly-ubiquitin chains with functionally distinct outcomes on Myc stability, i.e., degradation versus stability. The physiological relevance of regulating Myc activity through protein stability is underscored by observations that point mutations at or near T58, which render Myc resistant to proteasomal degradation, occur with high frequency in B-cell lymphomas [9].
Examples of viruses that modulate Myc activity include Kaposi's sarcoma associated herpesvirus (KSHV) and Epstein-Barr virus (EBV). Infection by these c-herpesviruses is characterized by the establishment of latent infection in memory B-cells. Access to this cell type is gained by virus-driven proliferation of germinal centre (GC) B-cells [10], where virus genomes replicate and segregate in step with normal cell division. This process is mediated by episomal maintenance proteins, which include EBV nuclear antigen-1 (EBNA-1) [11] and latency associated nuclear antigen (LANA) encoded by ORF73 of c-2-herpesviruses [12].
Given the essential role of Myc for the initiation and maintenance of GCs [13,14], it is not surprising that c-herpesviruses have evolved mechanisms to modulate Myc activity. In the case of KSHV-associated primary effusion lymphoma, Myc was shown to be abnormally stabilized [15,16]. The mechanism involves the direct interaction of the viral protein LANA with Gsk-3 resulting on reduced levels of Myc T58 phosphorylation [17]. Another strategy appears to be employed by EBV encoded EBNA-3C protein. This viral protein was shown to increase the transcriptional activity of Myc through an interaction with both Myc and SCF Skp2 . Surprisingly the mechanism proposed does not involve poly-ubiquitination but rather the action of SCF Skp2 functioning as a transcription co-factor for Myc [18].
Here, we utilized murid herpesvirus-4 (HuHV-4) infection of mice as a model system to address the role of Myc in the pathogenesis of c-herpesviruses. We show that Myc expression is required for the expansion of MuHV-4 infection in GC B-cells. The mechanism involves heterotypic poly-ubiquitination of Myc mediated through the ElonginC/Cullin5/SOCS (supressors of cytokine signaling) (EC 5 S) E3 ubiquitin-ligase activity of mLANA encoded by ORF73 of MuHV-4. EC 5 S mLANA mimics SCF b-TrCP by antagonizing SCF Fbw7 -mediated proteasomal turnover of Myc but unlike the cellular E3 ubiquitin-ligases its activity is not dependent on the phosphorylation status of Myc. Our results underscore the importance of modulating Myc activity during cherpesvirus driven lymphoproliferation providing a link between persistent infection and lymphoproliferative disease.

Myc transcriptional activity is up-regulated during MuHV-4 infection of GC B-cells
Experiments were designed to investigate the role of Myc expression on gammaherpesvirus pathogenesis. We utilized MuHV-4 infection of laboratory mice as the model, which is characterized by the expansion of latently infected B-cells in GCs and virus persistence in memory B-cells [19]. We analysed the transcription of several Myc target genes in infected versus noninfected GC B-cells, purified from the same pool of splenocytes. We utilized a recombinant MuHV-4 expressing a yellow fluorescent protein (YFP) [20] to segregate infected (CD19 + CD95 hi GL7 hi YFP + ) from non-infected (CD19 + CD95 hi GL7 hi YFP 2 ) GC B-cells derived from C57BL/6 mice, at day 13 post-infection. We analysed the transcription of Myc signature genes involved in cell cycle entry (encoding for cyclins B1, D1, D2 and E, and cyclin-dependent kinase 4) and GC B-cell activation (encoding for IL-10, B-ATF, MIF and CD70). Cyclin D3 was included as a gene non-regulated by Myc. The transcription of Myc target genes was significantly increased in infected GC B-cells when compared with their noninfected counterparts ( Figure 1A). These data show that during the expansion of latent infection in GC B-cells, MuHV-4 induces a transcription profile compatible with increased Myc transcriptional activity.

Myc is essential for GC responses
To define the impact of Myc expression on gamma-herpesvirus pathogenesis, we generated mice with conditional deletion of c-myc in B-cells undergoing GC reaction. This was achieved by breeding homozygous c-myc fl/fl mice, where second and third exons of the cmyc locus are flanked by two loxP sites [21] to heterozygous Cc1-cre mice, in which expression of Cre recombinase is induced by transcription of the Ig-c1 constant region gene segment early in GC development during immunoglobulin class-switch recombination [22]. Resulting progeny Cc1-cre KI/WT ;c-myc fl/fl , hereafter designated GC Myc KO, is expected to have specific deletion of cmyc in class-switched GC B-cells, after immunization with Tdependent antigens. Thus, we next investigated GC responses in the absence of Myc expression. GC Myc KO mice were immunized with the Th-2 cell-dependent antigen 4-hydroxy-3nitrophenylacetyl (NP)-chicken c-globulin (CGG) adsorbed to alum. Frequencies of GC B-cells (CD19 + CD95 hi GL7 hi ) were analysed at day 10 post-immunization. Compared to control litter mates Cc1-cre WT/WT ;c-myc fl/fl , hereafter designated control mice, GC Myc KO mice showed a marked impairment in GC development ( Figure 1B), accompanied by a strong reduction in IgG1 + B-cell numbers (CD19 + IgD 2 IgM 2 IgG1 + ) ( Figure 1C). Immunization with ovalbumin (OVA) emulsified in complete Freund's adjuvant (CFA) further confirmed that GC Myc KO mice were defective to mount a normal GC response, though less deficient than upon immunization with NP-CGG, and presented reduced levels of IgG1 expressing B-cells ( Figure 1D and 1E, respectively). However frequencies of IgG2a/2b + B-cells (CD19 + IgD 2 IgM 2 IgG2a/2b + ), whose class-switching is not strictly dependent on Ig-c1 promoter, revealed that GC Myc KO mice are competent in developing IgG2 class-switched B-cells ( Figure 1F). These data demonstrate a clear requirement for Myc expression in order to generate GC reactions to T cell dependent antigens. Our results are in direct agreement with those recently published that demonstrate the lack of GCs in mice in which c-myc is ablated early during GC induction [13]. Since we obtained the two transgenic mice lines carrying the alleles c-myc fl/fl and Cc1-cre from these authors, and independently generated GC Myc KO mice, our data are directly comparable.

Myc is critical for the establishment of persistent infection
The requirement of Myc for gammaherpesvirus pathogenesis was next investigated by infecting GC Myc KO mice with MuHV-4. Analysis of the percentages of GC B-cells revealed no significant differences between control and GC Myc KO mice (Figure 2A),

Author Summary
Being obligatory intracellular parasites, it is not surprising that viruses have evolved mechanisms to induce cellular proliferation to promote their own life cycle. This is notorious in the case of c-herpesviruses, such as Epstein-Barr virus (EBV) and Kaposi's sarcoma virus (KSHV), which are human pathogens associated with lymphoproliferative disease and several tumors. Host colonization by cherpesviruses is critically dependent on the ability to expand latent infection in proliferating B-cells. Virusinduced cellular proliferation is a process mediated by the expression of specific viral proteins. One of such proteins is the latency-associated protein (LANA) of KSHV. In this study, we use murid herpesvirus-4 (MuHV-4) as a mouse model of c-herpesvirus pathogenesis. We show that the MuHV-4 LANA (mLANA) stabilizes the cellular oncogene Myc, increasing its half-life, thus promoting its activity as a potent inducer of cellular proliferation. The molecular mechanism involves heterotypic poly-ubiquitination of Myc mediated via mLANA. The findings here reported demonstrate that modulation of Myc is essential for c-herpesvirus persistent infection, establishing a link between virus induced lymphoproliferation and disease. The implication is that revealing a critical function of a viral protein possibly allows the development of small molecule probes to disrupt mLANA-Myc interaction, therefore inhibit virus induced lyhophoproliferative disease.
with the majority of infected GC B-cells falling into dark zone ( Figure 2B) as previously described for infection of wild type mice [23]. To determine if MuHV-4 infected GC B-cells had been or not exposed to Cre-mediated c-myc deletion, we FACS purified infected cells from GC Myc KO mice, at day 14 post-infection. A PCR assay was employed to detect floxed and deleted c-myc alleles in DNA from infected GC B-cells, compared to DNA from uninfected total B-cells (CD19 + YFP 2 ), purified from the same pool  Figure 2C). This contrasted with total non-infected B-cell population where c-myc deletion could be readily detected ( Figure 2C). To further confirm the integrity of the c-myc locus in infected GC B-cells in GC Myc KO mice, we quantified Myc mRNA levels. Comparison of GC B-cells from GC Myc KO with wild type infected mice revealed no significant differences in Myc transcription ( Figure 2D). These data imply that MuHV-4 is expanding exclusively in GC B-cells where the c-myc locus did not undergo Cre-mediated deletion. Accordingly, GC Myc KO mice infected with MuHV-4 were unable to class-switch to IgG1, which

mLANA promotes Myc transcriptional activity and cell cycle progression as an EC 5 S ubiquitin-ligase
We have shown before that the ORF73 protein encoded by MuHV-4, designated mLANA by homology with the latency associated nuclear antigen encoded by KSHV, is selectively transcribed in GC B-cells [24]. Thus mLANA was a strong candidate to be responsible for the observed increased transcription of Myc target genes in MuHV-4 infected GC B-cells. To address this hypothesis, we analysed the transcription of Myc target genes in mLANA expressing cells. When compared to control transfected cells, mLANA expression induced the transcription of all Myc target genes analysed ( Figure 3A). Expression of mLANA had no effect on c-myc mRNA levels indicating that its putative modulatory effect on Myc was post-transcriptional ( Figure 3B). We have also shown before that mLANA acts as the substrate recognition factor of an ElonginC-Cullin5-SOCS (suppressor of cytokine signalling) (EC 5 S) E3 ubiquitin-ligase towards the p65/RelA cellular transcription factor NF-kB [25]. The mechanism involves the assembly of an EC 5 S -like complex, mediated by a viral unconventional SOCS-box motif present in mLANA. Hence, we analysed if mLANA-mediated modulation of transcription of Myc target genes could be attributed to its function as an E3 ubiquitin-ligase. To this end, we utilized a previously characterized mLANA mutant, designated mLANA-SOCS where residues V199, L202, P203 and P206 were substituted by alanines abrogating E3 ubiquitin-ligase function [25]. This mutant was no longer able to modulate the expression of Myc target genes when compared with intact mLANA ( Figure 3A). To define if the observed mLANA modulatory effect on cellular transcription was Myc specific, we carried out gene reporter assays using a synthetic Myc reporter plasmid containing three copies of E-box sequences driving the expression of luciferase. As expected, overexpression of Myc was translated into a significant increase on luciferase activity ( Figure 3C). Cells expressing mLANA exhibited comparable levels of luciferase activity, which increased further when Myc was concomitantly expressed. In contrast, mLANA-SOCS expression showed no effect on luciferase levels. We next proceeded to analyse the modulatory effect of mLANA on Myc in B-cells, which are physiological more relevant given the tropism of MuHV-4. When compared to control and mLANA-SOCS transfected cells, mLANA expression induced the transcription of all Myc target genes analysed ( Figure 3D). As before, expression of mLANA in A20 B-cells had no effect on c-myc mRNA levels confirming that its modulatory effect on Myc was post-transcriptional ( Figure 3E). Myc transcriptional activation of genes encoding proteins involved in cell cycle entry results in transition from G0-G1 to S phase. Thus, we analysed cell cycle profiles in B-cells expressing mLANA in comparison to control or mLANA-SOCS. These experiments showed a clear decrease in the number of mLANA expressing cells in G1 phase, with a concomitant increase in the number of cells in S and G2-M phases ( Figure 3F). Combined these data indicate that mLANA is modulating Myc-dependent transcription and progression through cell cycle in B-cells through its activity as an EC 5 S E3 ubiquitin-ligase. Co-immunoprecipitation experiments also showed that mLANA and Myc exist in the same heteromolecular complex in a context of virus infection. This was  demonstrated using a murine B-cell lymphoma-derived cell line latently infected with MuHV-4, designated S11 cells ( Figure 3G). We also showed that this interaction was reduced for mLANA-SOCS ( Figure 3H, compare lanes 3 and 5). This reduction could be due to lower levels of Myc in mLANA-SOCS expressing cells, when compared to mLANA, indicating that the interaction is independent of the SOCS-box motif. Alternatively it is plausible that the SOCS-box is participating in mLANA-Myc interaction.

mLANA mediates Myc poly-ubiquitination
We next set out experiments to investigate if EC 5 S mLANA was able to mediate poly-ubiquitination of Myc. We started by performing a nickel-nitrilotriacetic acid (Ni-NTA) pull-down in the presence of histidine-tagged ubiquitin. Upon culture and cell lysis, ubiquitinated proteins were extracted from total cellular lysates with Ni-NTA beads and resolved by SDS-PAGE. The levels of ubiquitinated Myc present in each condition were analysed by immunoblotting. We observed that when mLANA was expressed, the levels of ubiquitinated Myc were significantly enhanced ( Figure 3I, compare lanes 3 and 4). To confirm this activity in a more relevant biological context, we evaluated the ability of mLANA to promote the ubiquitination of endogenously expressed Myc. In comparison with control transfected and mLANA-SOCS transfected cells, higher levels of ubiquitinated Myc were detected in the presence of intact mLANA ( Figure 3J, compare lane 2 with 1 and 3). We have previously demonstrated that mLANA E3 ubiquitin-ligase activity towards p65/RelA required the E2 ubiquitin-conjugating enzyme UbcH5 [25]. Here we showed that EC 5 S mLANA requires UbcH5 as its E2 conjugating partner to mediate poly-ubiquitination of Myc ( Figure S1). Myc protein has 25 lysine residues that can be potentially ubiquitinated. To define if the ubiquitination ladder observed was due to the ability of mLANA to mediate poly-ubiquitination or multiple mono-ubiquitination of Myc in different lysine residues, we made use of a previously described lysine-free (K 2 ) version of Myc, which can only be ubiquitinated on its N-terminal residue [8]. By performing an in vivo ubiquitination assay with K 2 Myc, we observed the following. First, mLANA was able to mediate polyubiquitination of the N-terminal residue of Myc ( Figure 3K, compare lanes 3 and 4). Secondly, K 2 Myc in the presence of mLANA exhibited a ladder of ubiquitination indicating the assembly of poly-ubiquitin chains ( Figure 3K, compare lanes 3  and 4). Finally, the ubiquitination pattern of wild type Myc and K 2 Myc are distinctive, implicating that other lysine residue(s) are targets for mLANA-mediated poly-ubiquitination of Myc.

EC 5 S mLANA assembles heterotypic poly-ubiquitin chains on Myc
The addition of ubiquitin chains to a target protein can occur with different moieties that are emerging as determinants of biological outcome. These include, mono-ubiquitination and polyubiquitination. Ubiquin chains can be assembled using seven internal lysine residues (K) and thus define homotypic polyubiquitination (same K-linkage) or heterotypic poly-ubiquitination (mixed K-linkages) [26]. To investigate the type of lysine (K)linkage involved in mLANA-mediated poly-ubiquitination of Myc we utilized of a series of ubiquitin mutants with every single lysine residue, of the possible seven, substituted by an arginine. By performing in vivo ubiquitination assays in mLANA expressing cells we observed that in the absence of K33, K48 or K63 residues the ability of mLANA to promote Myc poly-ubiquitination was suppressed ( Figure 4A). Next, Myc transcription reporter assays were carried out in cells expressing the same ubiquitin mutants. Consistent with the poly-ubiquitination pattern, replacement of K33, K48 and K63 in ubiquitin rendered mLANA unable to positively modulate Myc transcriptional activity ( Figure 4B). Measurement of Myc cellular levels in extracts from mock transfected control or mLANA-transfected cells co-expressing each ubiquitin mutant, revealed that in the presence of wild type ubiquitin, mLANA-expressing cells exhibit augmented Myc levels ( Figure 4C, upper panel, compare lanes 1 and 2). The same result is observed when residues K6, K11, K27 and K29 of ubiquitin were substituted by arginines ( Figure 4C, upper and lower panels). However, in good agreement with the previous data, substitution of K33, K48 and K63, diminished or abolished the ability of mLANA to promote the increase in Myc cellular levels ( Figure 4C). Our observations demonstrate the involvement of different lysinelinkages in EC 5 S mLANA Myc poly-ubiquitination. However, these data do not distinguish between homotypic poly-ubiquitination at different K residues in Myc from mixed K-linkage polyubiquitination. Hence, we next utilized K 2 Myc where ubiquitination is only possible at the first metionine. In vivo ubiquitination assays and transcription reporter assays with K 2 Myc, in combination with each of the K ubiquitin mutants, essentially recapitulated the above observed results with wild type Myc ( Figure 4D and E). Hence, heterotypic poly-ubiquitination of the N-terminus of Myc is sufficient for mLANA modulatory activity. Collectively, these data show that EC 5 S mLANA requires different K-linkages to ubiquitinate Myc. Moreover, they demonstrate a direct correlation between the ability of mLANA to mediate Myc poly-ubiquitination, to increase Myc cellular levels, and to promote its transcriptional activity, supporting that all three activities are directly linked.

mLANA poly-ubiquitination on Myc is non-degradative
Although classically associated with protein degradation, ubiquitination is now emerging as regulator of a wide variety of non proteolytic cellular signalling functions [26]. Therefore, and in agreement with a positive effect on Myc activity, we further confirmed that mLANA-mediated poly-ubiquitination of Myc was non-degradative. To that end, we performed an in vivo ubiquitination assay in the presence of the proteasome inhibitor MG132. In control-transfected cells, the presence of MG132 favoured the accumulation of Myc ubiquitinated species ( Figure 5A, top panel, compare lanes 1 and 2), as well as the cellular levels of Myc protein ( Figure 5A, bottom panel, compare lanes 1 and 2). In contrast, in cells expressing mLANA, inhibition of proteasomal degradation had a negligible influence on both Myc poly-ubiquitination and Myc protein levels ( Figure 5A,  compare lanes 3 and 4). The importance of these results is threefold. Firstly, they confirm that under physiological conditions Myc turnover is highly regulated by the proteasome. Secondly, they show that treatment with the proteasome inhibitor MG132 has no effect on mLANA-mediated poly-ubiquitination of Myc, thus mLANA in not directing Myc for proteasomal degradation. Thirdly, the increase in Myc levels in response to mLANA expression was not altered under conditions of proteasomal inhibition, suggesting that expression of mLANA is preventing transfected to express K 2 Myc, the indicated ubiquitin mutants, mLANA (open bars) or control transfected (grey bars). Myc transcriptional activity was assayed as described in Figure 3C. Error bars represent SEM from triplicates from three independent transfection experiments. 2, without; +, with; a-, anti; IP, immunoprecipitation; TCL, total cellular lysates.  Collectively these results demonstrate that mLANA expression has a positive effect on Myc cellular levels by preventing its proteasomal turnover, thus prolonging its half-life, which is associated with increased transcriptional activity.

EC 5 S mLANA mimics SCF b-TrCP by antagonizing SCF Fbw7mediated proteasomal turnover of Myc
Under physiological conditions Myc half-life is tightly regulated through poly-ubiquitination by two distinct Skp1/Cul1/F-box (SCF) E3 ubiquitin-ligases. That is, poly-ubiquitination of Myc by SCF b-TrCP antagonizes SCF Fbw7 -mediated proteasomal dependent turnover [6,8]. Therefore, we hypothesised if mLANA could be modulating Myc stability by counteracting degradation by Fbw7 and mimicking the activity of b-TrCP. We overexpressed Fbw7 and analysed Myc cellular levels in the presence of co-expressed mLANA. As expected, overexpression of Fbw7 led to a decrease in Myc levels ( Figure 5C, compare lanes 1 and 2). Overexpression of Fbw7 had no effect on Myc levels when mLANA was concomitantly expressed ( Figure 5C, compare lanes 3 and 4). Notably, the antagonizing activity of mLANA towards Fbw7 was more pronounced when compared with that afforded by b-TrCP ( Figure 5C, compare lanes 4 and 6). We further characterized the mLANA effect on Fbw7 and b-TrCP interplay by depletion of the expression of the two cellular E3 ubiquitin-ligases and analysis of Myc levels. When Fbw7 was depleted Myc levels increased further by the presence of mLANA ( Figure 5D, compare lanes 3  and 4). In agreement, the turnover effect on Myc levels caused by depletion of b-TrCP was counteracted by concomitant expression of mLANA ( Figure 5E, compare lanes 3 and 4). Together these results demonstrate that mLANA mimics SCF b-TrCP by antagonizing SCF Fbw7 -mediated proteasomal turnover of Myc.

mLANA modulation of Myc activity is independent of the phosphorylation status of Myc
Fbw7 control of Myc turnover is dependent on the interaction between the ubiquitin-ligase and its substrate. Fbw7 recognizes Myc when phosphorylated on threonine (T) 58 and catalyses its poly-ubiquitination resulting in Myc proteasomal degradation [6,7]. Phosphorylation of Myc on T58 is sequentially preceded by phosphorylation on serine (S) 62, which activates and promotes Myc stability [5]. Thus T58 and S62 are key phospho-residues that regulate Myc activity at the protein level. Therefore, we set to investigate the influence of Myc phosphorylation on the modulatory activity of mLANA. Using phospho-specific antibodies we observed that under mLANA expression Myc is phosphorylated on both S62 and T58 ( Figure 6A, lane 2, first and second panels, respectively). To analyse if sequential phosphorylation of Myc was intact on mLANA expressing cells we proceed to substitute S62 or T58 to alanines (A) on Myc and assess the phosphorylation status of those Myc mutants. Compatible with the model of sequential phosphorylation, in which phosphorylation of S62 precedes phosphorylation of T58, MycT 58 A was readily phosphorylated on S62, whereas MycS 62 A was not phosphorylated ( Figure 6A, lanes 3-6, first and second panels, respectively). Remarkably, analysis of total Myc cellular levels revealed that, regardless of Myc phosphorylation on either S62 or T58, co-expression of mLANA led to increased Myc levels ( Figure 6A, third panel). These data not only demonstrate that mLANA is not interfering with Myc phosphorylation, but also indicates that mLANA modulatory functions override cellular pathways that control Myc activity. Consistent with this hypothesis, mLANA is co-immunoprecipitated by both Myc mutants ( Figure 6B), and it is able to mediate their poly-ubiquitination ( Figure 6C). Transcriptional activities of MycT 58 A and MycS 62 A were also increased by expression of mLANA ( Figure 6D), further supporting that mLANA targets Myc independently of cellular mechanisms of Myc regulation.

Discussion
In this study, we describe the first example of a viral protein activating Myc transcriptional activity through increased protein stability by mimicking a physiological post-translational regulatory pathway. Modulation of Myc function was shown to be essential for the lymphotropic c-herpesvirus MuHV-4 biology as infected cells exhibit increased expression of known Myc target-genes, and using a genetic approach, virus was found to amplify exclusively in intact Myc GC B-cells. The molecular mechanism involved heterotypic poly-ubiquitination of Myc mediated via the mLANA protein encoded by ORF73. This was reminiscent of a newly described pathway of Myc regulation through poly-ubiquitination. Popov et al. showed that the cellular E3 ubiquitin-ligase SCF b-TrCP uses UbcH5 ubiquitin-conjugating enzyme to form heterotypic poly-ubiquitin chains on the N-terminus of Myc. Poly-ubiquitination of Myc by SCF b-TrCP leads to Myc stabilization and was shown to antagonize SCF Fbw7 -mediated proteasomal turnover of Myc [8]. Like SCF b-TrCP , EC 5 S mLANA uses UbcH5 and antagonizes SCF Fbw7 . Furthermore, as previously demonstrated for SCF b-TrCP , single substitutions of K33, K48 or K63 of ubiquitin reduced or eliminated the ability of EC 5 S mLANA to polyubiquitinate, stabilize or increase the transcriptional activity of Myc. However, our results suggest that the molecular mechanism of mLANA modulation of Myc activity is not limited to Fbw7 antagonism. This is supported by the fact that mLANA protective effects on Myc stability, in conditions of Fbw7 over expression, are significantly more pronounced than observed for b-TrCP. Moreover, mLANA was able to increase the stability and activity of MycT58A, a Myc version that is not recognized by Fbw7. This contrasts with what has been reported for b-TrCP that albeit being able to poly-ubiquitinate MycT58A it did not impact on its turnover [8]. Thus, the ability of mLANA to increase Myc transcriptional activity through poly-ubiquitination, independently of the phosphorylation status of Myc on S58 and T62, indicates that this novel viral modulatory mechanism does not rely on posttranslation cellular regulation.
Modulation of host B-cell biology is of vital importance for cherpesviruses, as they depend critically on the expansion of latently infected B-cell in GC reactions for host colonization. GC reactions exhibit two distinct morphological areas. These include the dark zone (DZ), where centroblasts are rapidly dividing, and the light zone (LZ), where B-cells exit the cell cycle to differentiate into plasma cells or memory B-cells or reenter the DZ for additional rounds of cell division. Recently two studies have established the importance of Myc for GC biology [13,14]. These studies show that Myc expression is restricted to minute clusters of B-cells that initiate GCs and a small fraction of LZ B-cells. Furthermore, genetic interference with Myc expression or activity blocks GC formation and results in the collapse of mature GCs. Combined these studies demonstrate that expression of Myc is essential for the initiation and maintenance of GCs preceding B-cell proliferation in the DZ. It has been proposed that the lack of Myc expression in DZ B-cells, in conjunction with its short half-life, settles strict limits on the number of cell divisions afforded by centroblasts [27]. Therefore  is compatible with the selective expression of mLANA within GC B-cells [24] and its viral episomal maintenance properties [28].
By increasing the proliferative potential of latently infected GC B-cells through the modulation of Myc, c-herpesviruses are effectively promoting host colonization. However, different cherpesviruses have evolved distinct mechanisms to accomplish Myc modulation. KSHV achieves this via LANA through a mechanism that involves targeting of Gsk-3 [17] whereas EBV appears to increase the transcription activity of Myc via interaction of EBNA3C with SCF Skp2 [18]. However, how vital is Myc modulation for c-herpesvirus host colonization? We have previously reported that EC 5 S mLANA mediates poly-ubiquitinationdependent proteosomal degradation of the NF-kB family member p65/RelA [25]. In that study we demonstrate that a recombinant MuHV-4 with a disrupted SOCS-box motif in mLANA loses the ability of the virus to expand in GC B-cells and persist in the mouse. Given that this recombinant lacks ubiquitin-ligase activity we cannot ascribe its phenotype to NF-kB or Myc modulatory effects since both activities depend on an intact SOCS-box motif. We show here that mLANA interacts with Myc through a motif independent of the SOCS-box and have shown before that this also applies to interaction with p65/RelA [25]. We have mapped both interactions to the N-terminal half of mLANA but have been unable to identify any discrete binding motif to both cellular targets (unpublished observations). Since structural modeling of the N-terminal half of mLANA predicts it to be unstructured we envisage that conformational rather than linear binding motifs may be required for interaction of mLANA with p65/RelA and Myc. Current mLANA structural studies in our laboratory are addressing this question. However, the property of mLANA to modulate both Myc and p56/RelA supports that maintenance of a proliferative GC reaction through Myc stabilization requires simultaneous inhibition of NF-kB signaling.
The molecular basis for EC 5 S mLANA -mediated poly-ubiquitination of Myc and p65/RelA resulting in opposed outcomes is not known. Effector proteins with ubiquitin binding domains (UBDs) trigger specific cellular responses by recognizing different types of ubiquitin topologies [26]. Hence, the decoration of Myc and p65/ RelA with distinct K-linkages and lengths by EC 5 S mLANA may determine distinct ubiquitin-mediated cellular functions. It is interesting to note that in this respect a parallel exists between EC 5 S mLANA and SCF b-TrCP , as the latter also targets IkBa for proteasomal-mediated degradation [29] whereas it promotes Myc stabilization. The identification of UBD containing proteins that discriminate poly-ubiquitin chains topologies, which determine different biological outcomes is a field under intensive investigation. mLANA, therefore, provides as a good in vivo model for future studies.
Herein, we described a novel viral mechanism of stabilization of Myc through heterotypic poly-ubiquitination mediated by mLANA. The findings presented sustain the interpretation that increasing Myc stability is critical for the amplification of cherpesviruses in GC B-cells, thus persistence in the host. Therefore, this study provides a pathogenesis link between Myc and c-herpesviruses associated lymphoproliferative disease.

Reporter gene assays
For reporter gene assays, cells were transiently transfected with 500 ng of reporter vector, 1 mg of Myc and mLANA/mLANA-SOCS expression plasmids. In all transfections, a Renilla luciferase plasmid (10 ng) was used to normalise luciferase activity. Firefly and Renilla luciferase activities were assayed using Dual-Luciferase (Promega). Results are shown as fold induction relative to firefly luciferase activity measured in control-transfected cells.

Cell cycle analysis
Cell cycle distribution profiles were analysed using Vybrant DyeCycle Violet Stain (Invitrogen) according to the manufacter's instructions. Briefly, 24 h post-tranfection with GFP, GFP-mLANA or GFP-mLANA-SOCS expressing plasmids, 1610 6 A20 cells were incubated with 1 ml of Vybrant DyeCycle in complete RPMI for 30 minutes, at 37uC. The percentage of cells in the various phases of cell cycle was determined using FlowJo software (Tree Star, Inc), implementing the Dean-Jett-Fox model. Three independent experiments were performed for each experimental condition and a representative experiment is shown.

Transcription analysis of Myc target genes
Total RNA from FACS-purified uninfected or infected GC Bcells, or transfected HEK 293T cells or A20 B cells was extracted with Trizol (Invitrogen). RNA (500 ng) was used for cDNA synthesis (DyNAmo, Finnzymes). qPCR was performed using DyNAmo Flash SYBR Green (Finnzymes). Primer sequences are available in Table S1. All reactions were run in duplicates. Amplification efficiencies and threshold cycle values were defined by the fractional cycle number at which fluorescence crosses the fixed threshold. Relative mRNA values, normalized to GAPDH, were calculated by the Pfaffl method [31].

Laboratory animals
Cc1-Cre mice were provided by Dr. Kurosaki (Japan), with the agreement of Dr. Rajewsky and Dr. Casola. c-myc floxed mice were a gift from Dr. Moreno de Alborán (Spain). Cc1-cre KI/WT ; c-myc fl/fl mice were generated by breeding heterozygous Cc1-cre mice [22], with homozygous c-myc fl/fl mice [21]. Cc1-cre WT/WT ; c-myc fl/fl mice littermates were used as controls. C57BL/6 mice were obtained from Charles River Laboratories International Inc. Mice were bred and housed at IMM. All experimental protocols were performed in animals with 6-8 weeks of age.

Immunization and virus assays
Immunizations were performed via intraperitoneal injection with 100 mg NP-CGG (Biosearch Technologies, Inc.) adsorbed to 3 mg of aluminium hydroxide (SERVA Electrophoresis GmbH), or 100 mg ovalbumin (OVA) grade V in CFA (Sigma). OVA immunized mice were challenged 7 days post-primary immunization with the same antigen/adjuvant combination. Inoculation of MuHV-4 was performed intranasally with 10 4 p.f.u. in 20 ml of PBS under halothane. Frequencies of MuHV-4 genome-positive cells in GC B-cells were determined by limiting dilution combined with real-time PCR as previously described [32]. GC B cells were FACS purified from pools of five spleens using a BD FACSAria Flow Cytometer (BD Biosciences) and serially two-fold diluted. Eight replicates of each dilution were analysed by real time PCR (Rotor Gene 6000, Corbett Life Science). The primer/probe sets were specific for the MuHV-4 M9 gene (59 primer: GCCACGG-TGGCCCTCTA; 39 primer: CAGGCCTCCCTCCCTTTG; probe: 6-FAM-CTTCTGTTGATCTTCC-MGB). Samples were subjected to a melting step of 95uC for 10 min followed by 40 cycles of 15 s at 95uC and 1 min at 60uC. Real-time PCR data was analyzed on the Rotor Gene 6000 software. The purity of sorted cells was always greater than 97%, as analyzed by flow cytometry.
Plasmids, cell culture, DNA transfection and immunological reagents Information provided in protocols S1. Figure S1 UbcH5 is required for mLANA poly-ubiquitination of Myc. (A) mLANA immunoprecipitates exhibit E3 ubiquitin-ligase activity towards Myc, in vitro. Cell lysates from transiently transfected HEK 293T cells expressing mLANA or control transfected were subjected to immunoprecipitation with polyclonal anti-mLANA rabbit serum. After three washes in lysis buffer, immunoprecipitates were resuspended in reaction buffer (40 mM HEPES [pH 7.4], 60 mM potassium acetate, 1 mM EDTA, 2 mM DTT, 5 mM MgCl2, 10% glycerol). Myc was generated by transfection of 293T cells with a HA-tagged Myc expression plasmid, followed by immunoprecipitation with anti-HA antibodies after 48 hours of culture. After washing in lysis buffer, HA-Myc was eluted from beads using 0.5 mg/ml of HA peptide (Sigma). Reactions were supplemented with recombinant ubiquitin (2.5 mg) (Biomol International), E1 (50 ng), E2 (100 ng) (Calbiochem), GST-RelA (2.5 mg) and ATP regenerating buffer (Biomol International), when appropriate. Reactions were incubated for 1 hour at 30uC. Proteins were eluted in reduced Laemmli's buffer, resolved by SDS-PAGE and analyzed by immunoblotting with anti-ubiquitin antibody. (B) UbcH5 coimmunoprecipitates with mLANA dependent on mLANA SOCSbox motif. mLANA or mLANA-SOCS proteins were immunoprecipitated from total cellular lysates from HEK 293T cells transiently transfected with the expression plasmids (top). The presence of UbcH5 in the immunoprecipitates was analysed by immunoblotting. (C) mLANA-mediated activation of Myc is dependent on endogenous UbcH5 expression. HeLa cells were transfected with pSuper-puro vectors encoding control nontargeting (grey bars) or UbcH5 (open bars) directed shRNAs, along with the indicated expressing plasmids (bottom). Analysis of Myc transcriptional activity associated with each experimental condition was assessed as described in Figure 3C. Error bars represent the standard error of the mean from three independent experiments. 2, without; +, with; a-, anti; IP, immunoprecipitation; TCL, total cellular lysates. (TIFF)

Supporting Information
Protocol S1 Plasmids, cell culture, DNA transfection and immunological reagents. Description of plasmids and immunological reagents utilized. Cell culture and DNA transfections procedures are described.