Functional Characterization of the Interaction between Human La and Hepatitis B Virus RNA

RNA in living cells. In addition, we demonstrate in transient transfection experiments that disruption of the proposed La binding site diminishes HBV RNA half-life. Furthermore we show that recombinant hLa binds WT and mutant HBV RNA with the same affinity but with lower specificity arguing for the requirement of accessory factors for a specific interaction. Collectively, these results support the concept that hLa contributes to HBV RNA stability.


INTRODUCTION
RNA metabolism depends on the formation of ribonucleoprotein particles mediating diverse processes such as splicing, polyadenylation, nuclear export, and the regulation of mRNA stability (1,2). The formation of RNPs is a tightly controlled process, potentially regulated by several stimuli, including hormones and cytokines. Such stimuli can alter the RNA-binding activity of proteins on the post-translational level by phosphorylation or dephosphorylation and thereby the processing and stability of RNAs (3)(4)(5). In addition, RNA processing depends on various cis-acting elements including splice sites, export elements and endoribonucleolytic cleavage sites recognized by RNA binding proteins. In order to fully understand the regulation of processing of a specific RNA, both trans-acting factors and cisacting elements as well as their functions need to be known. The same applies for a detailed understanding of the metabolism of viral RNA. Such studies could lead to the identification of novel cellular targets valuable for the development of innovative antiviral strategies when focussed on the posttranscriptional control of RNAs of viruses with global medical importance. This applies to Hepatitis B Virus (HBV) with more than 300 million chronically infected carriers worldwide which await more effective antiviral therapies.
HBV is a noncytopathic, hepatotropic virus with a 3.2-kb circular DNA genome. After conversion into a cccDNA this genome serves in the nucleus as template for transcription of all viral RNAs. Synthesis of these transcripts is driven by at least four promoters leading to a large size heterogeneity with many different 5´-ends, while they all have very similar 3´-ends due to processing at the same polyadenylylation site (6). The so called pregenomic RNA (slightly longer than genome length) is encapsidated into nucleocapsids where it is reverse transcribed into viral DNA. This RNA serves also as messenger for synthesis of the viral P.protein as well as for the core protein, a regulator protein, designated as e-antigen. The viral surface proteins as well as a regulatory protein with a role in hepatocarcinogenesis, designated HBx, are translated from subgenomic mRNAs, 2.4 kb, 2.1 kb, and 0.7 kb in length.
of cytokines to induce the posttranscriptional downregulation of HBV RNA (for review see (7)) in a HBV transgenic mouse model (8), in human hepatoma cells (9,10) as well as the inhibition of duck HBV replication (11)(12)(13). In the HBV transgenic mouse model, the cytotoxic T lymphocyte (CTL) response to HBV antigens was shown to inhibit HBV replication by a posttranscriptional, non-cytotoxic mechanism leading to effective viral RNA degradation as induced by IFN-γ and TNF-α  (8,14,15). While trying to evaluate the intracellular mechanism(s) responsible for the cytokine-mediated posttranscriptional destabilization of HBV RNA, the mouse La autoantigen (homologue to human La autoantigen) has been identified. The La protein interacts with a small cis-acting element located within the viral RNA between position nt 1275-1291 (16,17). The tight temporal correlation between the cytokine-mediated down-regulation of HBV RNA and the cytokineinduced fragmentation of full-length La led to the assumption that full-length La stabilizes HBV RNA by interacting with the cis-acting element (18). Recently, it has been shown that HBV RNA is cleaved close to the La binding site by an endoribonucleolytic activity present in nuclear extracts prepared from HBV transgenic mice (19). Upregulation of this activity coincided with the cytokine-induced fragmentation of La and degradation of HBV RNA, supporting the assumption that HBV RNA is more accessible to endoribonucleolytic cleavage after disappearance of full-length La protein. More specifically, we hypothesize that La, in concert with additional trans-acting factors, forms an HBV RNA ribonucleoprotein complex stabilizing HBV RNA.
The La protein has been described as important cellular factor involved in the RNA metabolism of a variety of viruses. Most often a function of La was attributed to the translational regulation of viruses like polio and Hepatitis C (20)(21)(22)(23). In cultured cells relocalization of the predominantly nuclear La protein into the cytoplasm during infection was frequently observed (24,25). Moreover, La protein was reported to stabilize not only cellular histone mRNA (26), but also RNA of hepatitis C virus (27).
We studied whether hLa is associated with full-length HBV RNA, and evaluated the importance of the La binding site for the half-life of HBV RNA. We show that HBV RNA is co-precipitated with hLa, a strong indication for a physical interaction between hLa and HBV

EXPERIMENTAL PROCEDURES
Plasmid constructs and mutagenesis-The HBV expression plasmid pCH-9/3091 (kind gift of H. Schaller, Heidelberg, Germany) referred to as pHBV-WT contains a more than full-length HBV genome (subtype ayw) in which synthesis of the pregenomic RNA is under control of a CMV promotor. Mutations were introduced into the plasmid by PCR according to the site directed mutagenesis method (Stratagene, USA) using proof-reading Pwo DNA Polymerase  -CCA TAC TGC GGA AAT AAT AGC CGC TTG TTT   TGC TGC-3'), antisense primer (5'-CGA GCA AAA CAA GCG GCT ATT ATT TCC GCA   GTA TGG-3'). ptetHBV contains a more than full-length HBV genome (subtype ayw) in which synthesis of the pregenomic RNA is under control of the Tet promoter (kindly provided by Dr. Christoph Seeger, (28)). To introduce mutation M-2 into ptetHBV plasmid, the Dra III and Nco I fragment of pHBV-M2 was ligated into the Dra III and Nco I linearized ptetHBV using standard methods. The plasmid was designated ptetHBV-M2. Correct mutagenesis was confirmed by sequencing using the Taq-cycle sequencing protocol with IRD-800 labeled primers (MWG Biotech, Ebersberg, Germany) and a Licor automated sequencing device (MWG Biotech). ACG C-3´). Plasmid pEGFP-N1 was used for the production of DNA templates for the generation of GFP antisense transcripts. Two primers were used: the antisense primer (5'-gga tcc taa tac gac tca cta tag gGT CCA TGC CGA GAG TGA TCC C-3') contained a restriction site for ClaI (shown in italic), the T7 RNA polymerase promoter sequence (shown in bold), the GFP sequence (capitol letters) and the sense primer contained GFP sequences (5´-CCT GGT CGA GCT GGA CGG C-3´). PCRs for the templates were performed with 1 ng plasmid and the mixture contained 80 pmol of each primer in 1 X PCR buffer, 0.2 mM of GTP, ATP, TTP and CTP, and 2.5 U of Taq DNA polymerase (Roche). PCR was performed as follows: 5 min at 95°C, followed by 35 cycles of 1 min at 95°C, 1 min at 56°C, 1 min at 72°C and finally once 5 min 72°C. The PCR products were purified by size exclusion using Northern Blot Analysis-HuH7 cells were harvested 48 h or as indicated after transfection.
Total RNA was prepared using TriPure Isolation Reagent (Roche) according to the manufacturer´s protocol. 10 µg of total RNA was separated on a 1.2 % agarose-formaldehyde gel. RNA was blotted onto a nylon membrane (Osmonics, Westborough, USA) and hybridized with 32 P-labeled in vitro transcribed RNA probes overnight at 68°C. To standardize transfection efficiency and RNA loading, Northern blots were hybridized with 32 P-labeled in vitro transcribed antisense GFP probe and antisense glycerinaldehyde-3phosphate dehydrogenase (GAPDH) or antisense histone 2A probes, respectively. Blots were exposed to Fuji imaging screens, and signals were quantified by Fujix BAS 2000 bio-imaging analyzer (Fuji, Japan) and by TINA software (Raytest, Germany). For northern blot analysis the complete RNA pellet was solved in 40 µl loading buffer and loaded onto a 1.2 % agarose-formaldehyde gel and processed as described above. Western blot analysis was performed by standard methods, using 12.5 % polyacrylamig gels and nitrocellulose membranes.
For RT PCR analysis the RNA pellets (starting material, pellets , supernatants) were solved in 43 µl DEPC treated water, subsequently 5 µl RNase free DNase reaction buffer (Promega) and 2 units RNAse free DNAse (Promega) were added and incubated for 30 min at 37 °C.
DNAse was inactivated by heating the sample for 10 min at 70°C. 2 µl of DNase treated RNA was used as template for RT PCR reaction. RT PCR was conducted using the Titan One-Tube RT PCR kit (Roche) and following primer pairs were used: HBV Sense 5´-

RESULTS
The hLa protein is associated with HBV RNA-Recently it has been shown that the cytokineinduced posttranscriptional degradation of HBV RNA tightly correlates with the cytokinedependent processing of the mouse La protein (18). In addition it has been shown that recombinant hLa interacts with in vitro transcribed HBV RNA.B (17). These data suggest an association of La and HBV RNA also cell culture. Accordingly, we asked whether it would be possible to co-precipitate HBV RNA with hLa specific antibodies for immunoprecipitations (IP). For this purpose HepG2 2.15 cells, stable expressing HBV RNA (31), were harvested, lysed, and subsequently the cleared cell lysate was incubated with either anti-La antibodies or mouse IgG 2A coated protein A sepharose beads. As shown in figure 1A and B hLa was specifically precipitated and not detectable in the pellet of the control IP. To detect viral RNA in the IPs pellets, RNA was extracted from additional pellets immunoprecipitated with hLa specific as well as unspecific antibodies and analyzed by RT-PCR ( Fig. 1 A) and northern blotting ( Fig. 1 B). RT-PCR was performed with specific primers for HBV RNA, TOP RNA L-37 and the hnRNP E2 RNA. Recently the coprecipitation of L-37 mRNA but not of the hnRNP E2 with La was shown (32) and were used as positive and negative controls in our study, respectively. As shown by RT-PCR HBV RNA and L-37 mRNA were specifically amplified in the La immunoprecipitation pellet but not in the IgG control pellet. In contrast hnRNP E2 RNA was not co-precipitated. This analysis strongly suggests an association between hLa and HBV RNA. To further verify the association between hLa and HBV RNA, we performed an additional Coimmunoprecipitation experiment (Fig. 1 B) and analyzed the RNA extracted from the IP pellets for HBV RNA by Northern blot analysis. This experiment reveals that full-length HBV pregenomic RNA was co precipitated with hLa but not with IgG´s although to a very small extent. Taken together both experiments clearly show the association between hLa and HBV RNA and suggest a physical contact between hLa and HBV RNA in living cells.
Mutagenesis of the La binding site reduces HBV RNA levels-It has been reported earlier that mouse La present in nuclear extracts binds a predicted stem loop structure with high affinity (approx. K D of 1 nmol) and with high specificity (16) in vitro. In contrast highly purified recombinant hLa binds HBV RNA.B with high affinity but with low specificity (17). compared to HBV RNA levels transcribed from pHBV-WT (Fig. 3A, lanes 1, 2 versus 3, 4). Mutagenesis of the La binding site reduces HBV RNA half-life-Next we asked whether reduced HBV RNA-M2 levels compared to HBV wt RNA were due to an accelerated turn over rate. Therefore, the half-life of wild type and pHBV-M2 pregenomic RNA was determined using a tetracycline (tet) controlled HBV expression system. Mutation M2 was introduced into the tetracyclin-regulated HBV expression vector (ptetHBV-WT, (28)) referred to as ptetHBV-M2. This system allows the specific shut-off of pregenomic HBV RNA transcription without affecting the general cellular transcription machinery which would be blocked by the use of inhibitors such as actinomycin D. Hence, potential ambiguous results due to secondary effects caused by inhibiting general transcription were prevented. In this system synthesis of the pregenomic RNA was under the control of a CMV promoter inactivated after addition of the tetracycline analog doxycycline. Huh-7 cells were transiently transfected with ptetHBV-WT and ptetHBV-M2 plasmids, and HBV RNA halflife was determined by Northern blot analysis. A substantial amount of pregenomic RNA levels was raised 24 hours after transfection at which time doxycycline was added to abolish further HBV RNA synthesis (Fig. 4A, lanes 0h). Three, six and eight hours thereafter cells were harvested, RNA prepared and analyzed. Fig. 4A shows a representative Northern blot analysis of the half-life determination of pregenomic RNA. As expected from the reduced HBV-M2 RNA levels observed in our previous experiments, the pregenomic RNA levels of WT HBV were higher than that of HBV-M2 24 h post transfection immediately before doxycyclin addition to the cell culture medium (Fig. 4A, lanes 1, 2 versus 9, 10). Three hours after doxycyline treatment, pregenomic RNA levels were decreased to about 30% for WT HBV and to about 50% for HBV M2 (Fig. 4A, lanes 0h versus lanes 3h; and Fig. 4B). The pregenomic RNA levels decrease further to about 30% and 20% during the following observation period up to 8h post addition of doxycyclin (Fig. 4A, lanes 6h and 8h; and and -M2 (Fig. 5A, lanes 7-9) were similarly efficient in competing for the binding of

Modulation of the hLa-HBV RNA.B interaction by nuclear factors and phosphorylation-To
test whether additional cellular factors modulate the interaction between highly purified recombinant hLa and HBV RNA.B we analyzed the binding in the presence of nuclear extracts prepared from Huh7 cells. Different amounts of recombinant hLa or nuclear extracts were incubated with labeled HBV RNA.B-WT and analysed by EMSA. As shown in Figure   6, 50 and 100 ng recombinant hLa without nuclear extract forms two major complexes with HBV RNA.B indicated as monomers and multimers (Fig. 6A, lanes 2-7). Next different amounts of nuclear extracts were tested revealing the formation of a major complex of low mobility (indicated by a arrow heads, Fig. 6B, lanes 2-7) and a minor complex of higher mobility (indicated by a star, Fig. 6B, lanes 2-7). The nature of these complexes is unknown  7 versus 12, 13). In contrast an unrelated antibody was unable to shift the hLa/HBV RNA.B complexe (Fig. 6C, lanes 14, 15). These data demonstrate that components present in nuclear extracts modulate the binding behavior by preferentially supporting the formation of monomeric hLa HBV RNA.B RNPs. These findings indicate that additional factors are involved in specific fromation of hLa/HBV RNA RNPs.
As shown earlier, dephosphorylation of mouse nuclear extracts strongly diminished the ability of mouse La to bind to HBV RNA.B (16). In that study it was not possible to discriminate whether the complex formation was abolished due to the dephosphorylation of endogenous mouse La or of accessory factors. La is mainly phosphorylated by casein kinase II at serine 366, and it was shown earlier that this modification did not alter the binding of hY1-RNA (33). We now tested whether phosphorylation of recombinant hLa by casein kinase II or dephosphorylation of hLa by alkaline phosphatase affected binding of HBV RNA.B. As shown in Fig. 7A  probably structural rather than sequential features of the RNA element are important to uphold HBV RNA half-life. In this context it is important to note, that the sequence of the stem-loop is highly conserved between a diversity of HBV genotypes and isolates (Fig. 1, B).
Our study raises the interesting question whether the nucleotide substitutions within the This view concurs with current accepted pathways of mRNA degradation. In general it is believed that the fate of an mRNA is regulated by trans-acting factors interacting with cisacting elements within an RNA molecule thereby protecting or destabilizing the RNA (for review (36)). Among the best characterized RNA destabilizing elements are AU-rich regions first described to regulate mRNA stability of short lived RNAs (37). The AU-rich regions are recognized by a group of proteins called AU-rich binding proteins (38). In most cases these proteins destabilize RNAs by interacting with the AU-rich regions, but stabilizing effects have also been reported (39)(40)(41). To our knowledge, it is not known how exactly this interaction destabilizes/stabilizes RNA. The stable alpha-globin mRNA has been described to interact with several proteins forming the so called alpha-complex associated with the 3'- Previously it has been shown that binding of mouse La to HBV RNA.B in vitro was abolished after dephosphorylation of the La containing nuclear extract (16). In this previous study, it was not possible to distinguish, whether impaired RNA binding activity was due to dephosphorylation of the phosphoprotein La or of other factors present in the nuclear extracts. Our observation that phosphorylation or dephosphorylation of recombinant hLa had no effect on binding HBV RNA.B, whereas dephosphorylation of nuclear extracts partially inactivated the activity promoting formation of the monomeric La complex showed that a phosphorylation-dependent activity modulates hLa binding to HBV RNA.B. We believe that this initial data support our view that the binding specificity of hLa to HBV RNA.B is mediated by phosphorylation-dependent host factors. In the absence of those factors recombinant hLa binds HBV RNA.B with low specificity. In this context it is important to note that the newly identified hLa interacting protein is also a phosphoprotein and that it will of interest to study whether the interaction between hLa and this protein might be regulated by phosphorylation. Furthermore, it will be interesting to find out which cellular signaling pathways affect the proposed phosphorylation-dependent composition of HBV RNPs thereby potentially regulating the turnover of viral RNAs. Importantly, it has been shown recently that janus kinase activity is required for the antiviral affect of interferons on HBV replication (35), indicating that cell signaling pathways and thereby phosphorylation/dephosphorylation events are required. This might also applies for cytokine-induced degradation of HBV RNA decay as supported by our finding that binding of mouse and human La to HBV RNA is modulated by accessory factors in a phosphorylation-dependent manner. Previously it has been shown that a variety of cellular factors involved in the regulation of mRNA stability, like RNA binding proteins were regulated by phosphorylation/dephosphorylation (50,51) and that ribonucleases activities were elevated by hormones (48,52,53), indicating that posttranslational modification and signal transduction pathways are important cellular mechanism determining mRNA levels.As a model, we assume that cytokine-induced HBV RNA degradation is initiated at the La binding site and this concept is supported by the following findings. First, the cytokine-induced HBV RNA degradation occurring in the mouse system in parallel with La processing is associated with increased endoribonuclease activity which leads to cleavage of the HBV RNA close to the La binding site (18,19). inducing structural changes and/or exposing the cleavage site located close to the predicted stem loop to endoribonuclease (19), thereby initiating HBV RNA degradation.
We conclude that the La binding site together with the cleavage sites compose a destabilizing element leading to HBV RNA degradation under specific cellular conditions. It will be of major importance to identify host factors interacting with hLa and/or HBV RNA in a phosphorylation-dependent manner to understand in more detail the initiation of HBV RNA degradation.   Mutation/sequence indicates in how many of the sequences analyzed the indicated subsituion was found C) Outline of nucleotide substitution introduced into the predicted stem loop region. Nucleotide substitutions are shown in bold.