Report

The Kaposin B Protein of KSHV Activates the p38/MK2 Pathway and Stabilizes Cytokine mRNAs

See allHide authors and affiliations

Science  04 Feb 2005:
Vol. 307, Issue 5710, pp. 739-741
DOI: 10.1126/science.1105779

Abstract

Cytokine production plays a critical role in diseases caused by Kaposi's sarcoma–associated herpesvirus (KSHV). Here we show that a latent KSHV gene product, kaposin B, increases the expression of cytokines by blocking the degradation of their messenger RNAs (mRNAs). Cytokine transcripts are normally unstable because they contain AU-rich elements (AREs) in their 3′ noncoding regions that target them for degradation. Kaposin B reverses this instability by binding to and activating the kinase MK2, a target of the p38 mitogen-activated protein kinase signaling pathway and a known inhibitor of ARE-mRNA decay. These findings define an important mechanism linking latent KSHV infection to cytokine production, and also illustrate a distinctive mode by which viruses can selectively modulate mRNA turnover.

Kaposi's sarcoma–associated herpesvirus (KSHV) is a γ-2 herpesvirus etiologically linked to Kaposi's sarcoma (KS) (1), an unusual tumor composed of proliferating, spindle-shaped endothelial cells, slit-like neovascular spaces, and inflammatory cell infiltration (2). KS spindle cells, which are latently infected with KSHV, elaborate a variety of pathogenetically important proinflammatory cytokines and angiogenic factors (3, 4); in turn, cultured spindle cells require cytokines [including interleukin-6 (IL-6)] for their survival and proliferation (5). KSHV also causes several rare lymphoproliferative diseases (6, 7), and these too are associated with elevated production of cytokines, notably IL-6 (610). However, the mechanisms linking KSHV and host cytokine production have remained obscure.

Cells latently infected by KSHV express only a handful of viral genes, including the products of the kaposin locus (11) (Fig. 1A). This locus contains a small coding region [open reading frame (ORF) K12] preceded by two families of 23-nucleotide GC-rich direct repeats (termed DR1 and DR2), and is transcribed as a single mRNA encompassing all three components. A complex translational program generates a variety of proteins from this mRNA (12). Kaposin B, the subject of this report, results from translation of the repeats alone and consists of a series of tandemly repeated copies of 23–amino acid peptides derived from translation of the DR2 (HPRNPARRTPGTRRGAPQEPGAA) and DR1 (PGTWCPPPREPGALLPGNLVPSS) repeats (12, 13). Its function has heretofore been unknown.

Fig. 1.

(A) Schematic map of the K12 (kaposin) locus derived from a pulmonary KS isolate of KSHV (12). (B) Coimmunoprecipitation of kaposin B and MK2. 293T cells were cotransfected with expression vectors encoding MK2 and FLAG-tagged kaposin B. Cell lysates were immunoblotted with polyclonal antibody to MK2 (anti-MK2) (upper panel). In parallel, portions of the same lysates were immunoprecipitated with FLAG monoclonal antibody (mAb) and immunoblotted with anti-MK2 (lower panel). (C) Recombinant full-length MK2 purified from E. coli was incubated with the indicated GST fusion proteins bound to glutathione-Sepharose beads. MK2 was detected by Western blotting with anti-MK2. (D) Kaposin B binds to a central portion of the C-lobe of MK2. Truncated versions of MK2 were labeled with [35S]methionine by in vitro translation and tested for binding to GST–kaposin B (see fig. S1 for details). NLS, nuclear localization signal; NES, nuclear export signal. (E) Coexport of kaposin B and MK2 in response to stress signals. SLK endothelial cells were cotransfected with plasmids encoding kaposin B and green fluorescent protein–MK2 fusion protein. After 30 hours, cells were treated with LPS (1 μg/ml) for 30 min, then fixed and stained with kaposin B mAb and rhodamine-conjugated secondary antibody. DAPI, 4′,6′-diamidino-2-phenylindole.

To identify proteins that interact with kaposin B, we used the DR2 and DR1 repeats to screen a library of infected cell cDNAs in the yeast two-hybrid system (14). This yielded a clone consisting of amino acids 188 to 400 of MK2 [mitogen-activated protein kinase (MAPK)–associated protein kinase 2], a protein kinase whose activity is controlled by phosphorylation by p38, a MAPK activated by inflammatory and other signals (15). Immunoprecipitation of FLAG epitope–tagged kaposin B results in efficient coprecipitation of cotransfected MK2 in 293T cells, confirming the interaction in vivo (Fig. 1B, bottom panel, lane 4). This interaction is direct: A purified, Escherichia coli–derived fusion protein of glutathione S-transferase (GST) and kaposin B, but not GST alone, efficiently binds to purified recombinant MK2 (Fig. 1C). Parallel experiments with GST-DR1 and GST-DR2 proteins mapped the interaction domain to the DR2 region (Fig. 1C). Using truncated MK2 proteins, we also mapped the domains of MK2 required for interaction with kaposin B. Kaposin B binds to a region in the C-lobe of the kinase domain of MK2, with residues 200 to 270 apparently critical for the interaction (Fig. 1D) (fig. S1). This region contains a highly conserved activation segment, including a site for phosphorylation by p38 MAPK (Thr222) and a binding site for the autoinhibitory C-terminal domain of MK2 (16, 17).

These in vitro binding findings are further corroborated by studies that examine the colocalization of the two proteins under various conditions in vivo. Expression of kaposin B in cells reveals that the protein is colocalized with MK2 in the nucleus (Fig. 1E), despite the absence of a classical nuclear localization signal in kaposin B. MK2 is known to be exported from the nucleus in response to inflammatory signals that activate the p38 MAPK pathway [e.g., tumor necrosis factor–α (TNF-α) and lipopolysaccharide (LPS)] (18). It is note-worthy that treatment of cells with LPS (Fig. 1E) or TNF (fig. S2) causes kaposin B and MK2 to be coexported from nucleus to cytosol.

Binding of kaposin B to MK2 leads to activation of MK2, by several criteria. First, cells expressing kaposin B displayed higher levels of phosphorylated (activated) MK2 as judged by immunoblotting (Fig. 2A). Second, relative to MK2 precipitated from control cells, endogenous MK2 immunoprecipitated from cells expressing kaposin B was more active in phosphorylating the MK2 substrate heat shock protein 27 (Hsp27) (Fig. 2B).

Fig. 2.

Kaposin B stimulates MK2 kinase activity and blocks the degradation of cytokine mRNAs. (A) HeLa cells were transfected with empty vector or kaposin B expression vector, lysed, and harvested. Portions of the whole-cell lysates were immunoblotted with antibodies to MK2 (lower panel) and phosphorylated MK2 (upper panel). (B) Portions of the same HeLa cell lysates were immunoprecipitated with anti-MK2, and kinase activity was measured in vitro with recombinant Hsp27 as a substrate. (C) HeLa–Tet Off cells were cotransfected with β-globin–based reporter and test plasmids. After 30 hours, doxycycline was added to the media to stop transcription. RNA was harvested at 0, 1, 2, and 4 hours after doxycycline addition; β-globin mRNA was detected with a 32P-labeled antisense riboprobe. (D) HeLa–Tet Off cells were transfected with a β-globin–ARE expression vector and either an empty vector or kaposin B expression vector. One hour before addition of doxycycline, 1 mM hydrogen peroxide was added to one vector-transfected culture (top panel). Samples of RNA were Northern blotted for β-globin mRNA as above. (E) Human foreskin fibroblasts were transfected with empty vector or kaposin B expression vector. After 48 hours, cell supernatants were collected and extracellular GM-CSF and IL-6 levels were measured by ELISA.

Elevated MK2 kinase activity has a number of functional consequences, the best characterized of which is the stabilization of cytokine transcripts and other mRNAs that contain AU-rich elements (AREs). AREs typically consist of multiple copies of the sequence AUUUA, usually located in the 3′ untranslated region (3′UTR) of the transcript. Their presence leads to drastic destabilization of the mRNA in the ground state (1921), but this instability can be abrogated by activation of MK2 (22). Accordingly, we tested the effects of kaposin B expression on this activity with the use of a standard assay of ARE-mediated mRNA decay (23) based on a globin reporter gene. Fig. 2C shows that the β-globin mRNA reporter is quite stable and is unaffected by cotransfection of kaposin B. However, the insertion of an ARE [here, from the 3′UTR of granulocyte-macrophage colony-stimulating factor (GM-CSF)] into the β-globin gene renders the chimeric mRNA susceptible to rapid degradation, and cotransfection of kaposin B with the β-globin–ARE vector strikingly reverses this effect. The magnitude of this stabilization is similar to that induced by other, more potent, triggers of the p38-MK2 pathway, including oxidative stress (Fig. 2D). Correspondingly, transfection of kaposin B into cells resulted in major augmentation of both GM-CSF and IL-6 production, as determined by enzyme-linked immunosorbent assay (ELISA) of the culture medium 48 hours after transfection (Fig. 2E) (fig. S3).

To determine whether ARE-mediated mRNA decay is also impaired in KSHV latency, we transfected HeLa–Tet Off cells with a β-globin–ARE expression vector, then infected the cells with KSHV virions; this resulted in a latent KSHV infection. Doxycycline was then added to arrest transcription, and β-globin mRNA levels were determined at 0, 1, 2, and 4 hours thereafter. Latently infected cells displayed substantial prolongation of the half-life of the ARE-containing transcript (Fig. 3, A and B). To establish that this effect was due to kaposin B and not to other latency proteins, we tested the known latent proteins for their ability to block ARE-mediated mRNA decay in transfected HeLa–Tet Off cells. None of the other known latency genes—those encoding latency-associated nuclear antigen (LANA) (orf 73), v–cyclin D (orf72), v-FLIP (orf71), kaposin A (K12), or interferon regulatory factor v-IRF3—were active in ARE-dependent mRNA stabilization (Fig. 3C). Correspondingly, the only latent gene product capable of activating MK2 kinase activity in such transfectants was kaposin B (Fig. 3D).

Fig. 3.

Latent KSHV infection blocks degradation of ARE-containing mRNAs. (A) HeLa–Tet Off cells were transfected with β-globin–ARE reporter 24 hours before infection with KSHV. At 24 hours after infection, doxycycline was added to the media to stop transcription. RNA was harvested at 0, 1, 2, and 4 hours after doxycycline addition; β-globin mRNA was detected with a 32P-labeled antisense riboprobe. (B) Phosphorimager-based quantitation of the results. Open circles, GM-CSF ARE + vector; solid circles, GM-CSF ARE kaposin B; open squares, IL-6 ARE + vector; solid squares, +IL-6 ARE + kaposin B. (C) Other KSHV latency genes do not block ARE-mediated mRNA decay. HeLa–Tet Off cells were cotransfected with the indicated β-globin–ARE reporter and test plasmid. After 30 hours, doxycycline was added and RNA was harvested and Northern blotted as above. (D) HeLa cells were transfected with plasmids encoding kaposin B or other latency-associated KSHV genes. At 30 hours after transfection, cells were lysed and immunoprecipitated with anti-MK2. MK2 kinase activity was assayed with Hsp27 substrate and an antibody to phosphorylated Hsp27 (top panel). Total immunoprecipitated MK2 was measured with anti-MK2 (bottom panel).

How does kaposin B expression lead to MK2 activation? Although our understanding of this process is still incomplete, important clues emerged when we examined the state of activation of the other components of the p38 MAPK pathway. HeLa cells transfected with kaposin B display activation of p38 activity as well as that of MK2 (Fig. 4A). In fact, both p38 and MKK6, an upstream kinase implicated in p38 activation, are coimmunoprecipitated from cell lysates along with kaposin B–MK2 complexes (Fig. 4B). Although MK2 binds very efficiently to kaposin B (Figs. 1C and 4C), to date we have detected only very weak binding of purified p38 to recombinant kaposin B in vitro (Fig. 4C). This suggests that p38 is recruited to kaposin-MK2 complexes principally (though perhaps not exclusively) via its binding to MK2. The enhanced p38 activity observed in kaposin B–transfected cells is an important contributor to the net state of MK2 activation in such cells, because blockade of p38 activity with the selective inhibitor SB203580 results in substantial (although not complete) reduction of MK2 activation (Fig. 4D) and cytokine release (fig. S4).

Fig. 4.

p38 participates in kaposin B–MK2 signaling complexes. (A) HeLa cells were transfected with empty vector or kaposin B expression vector for 30 hours, lysed, and harvested. Portions of the whole-cell lysates were immunoblotted with antibodies to p38 (middle panel) and dual-phosphorylated p38 (upper panel). Dual-phosphorylated p38 was immunoprecipitated from these lysates, and p38 activity was measured with the use of activating transcription factor 2 (ATF2) as a substrate and immunoblotting with antibody to phosphorylated ATF2 (lower panel). (B) MKK6 and p38 coimmunoprecipitate with kaposin B–MK2 complexes. 293T cells were transfected with the indicated combinations of hemagglutinin (HA) epitope–tagged kaposin B, MK2, p38, and MKK6 for 48 hours, lysed, and immunoprecipitated with HA mAb. Western blots of the immunoprecipitated material and the whole-cell lysates were probed with polyclonal antibodies to MKK6, p38, MK2, and HA. (C) p38 and MK2 proteins (purified from E. coli) were incubated overnight at 4°C with the indicated GST-fusion proteins bound to glutathione-Sepharose beads, washed, electrophoresed, and immunoblotted with specific polyclonal antibodies. (D) Inhibition of p38 partially blocks kaposin B–mediated increases in MK2 activity. The selective p38 inhibitor SB203580 was added to kaposin B and empty vector transfected HeLa cells for 1 hour, cells were lysed, and whole-cell lysates were immunoblotted with the indicated antibodies. Portions of these lysates were immunoprecipitated with anti-MK2 and assayed for MK2 kinase activity. (E) Proposed model of MK2 activation by kaposin B. The reiterated DR2 repeats of kaposin B may bind multiple MK2 proteins in the nucleus, allowing for efficient phosphorylation by p38 MAPK. Phosphorylation of MK2 target proteins results in a blockade in ARE-mediated mRNA degradation, leading to enhanced production of proteins from ARE-containing transcripts. These proteins include cytokines as well as signaling molecules such as MKK6, which in turn could amplify p38 MAPK activation.

These results are consistent with a model (Fig. 4E) in which initial MK2 activation after its binding by kaposin B is amplified by p38 activation, leading to further activation of MK2. Much remains to be learned about the mechanisms of MK2 and p38 activation by kaposin B. Binding of MK2 by kaposin B could lead directly to its activation through an induced conformational change; we note that the region on MK2 to which kaposin B binds (Fig. 1D) is consistent with activation via displacement of the inhibitory C-terminal autoregulatory domain of MK2. Bound kaposin B could also shield activated MK2 from the action of cellular phosphatases. As to how p38 becomes activated, because cytokines such as IL-6 can activate the p38 pathway, elevated cytokine release from kaposin B–expressing cells may promote an autocrine or paracrine amplification loop that further enhances p38 activity. Irrespective of its mechanistic details, the finding that kaposin B activates the p38-MK2 pathway forges an important biochemical link between KSHV infection and the enhanced cytokine production that characterizes so many of its associated disease states, and provides a striking example of virus-mediated modulation of mRNA stability.

Supporting Online Material

www.sciencemag.org/cgi/content/full/307/5710/739/DC1

Materials and Methods

Figs. S1 to S4

References and Notes

View Abstract

Navigate This Article