ADAR1 enhances HTLV-1 and HTLV-2 replication through inhibition of PKR activity

Background The role of innate immunity in general and of type I interferon (IFN-I) in particular in HTLV-1 pathogenesis is still a matter of debate. ADAR1-p150 is an Interferon Stimulated Gene (ISG) induced by IFN-I that can edit viral RNAs. We therefore investigated whether it could play the role of an anti-HTLV factor. Results We demonstrate here that ADAR1 is also expressed in the absence of IFN stimulation in activated primary T-lymphocytes that are the natural target of this virus and in HTLV-1 or HTLV-2 chronically infected T-cells. ADAR1 expression is also increased in primary lymphocytes obtained from HTLV-1 infected individuals. We show that ADAR1 enhances HTLV-1 and HTLV-2 infection in T-lymphocytes and that this proviral effect is independent from its editing activity. ADAR1 expression suppresses IFN-α inhibitory effect on HTLV-1 and HTLV-2 and acts through the repression of PKR phosphorylation. Discussion This study demonstrates that two interferon stimulated genes, i.e. PKR and ADAR1 have opposite effects on HTLV replication in vivo. The balanced expression of those proteins could determine the fate of the viral cycle in the course of infection. Electronic supplementary material The online version of this article (doi:10.1186/s12977-014-0093-9) contains supplementary material, which is available to authorized users.

Innate immunity plays a critical role in host response to a viral infection by triggering effector mechanisms that restrict infection [21]. Effectors of innate immunity such as restriction factors are present in cells before infection takes place and can act as immediate inhibitors of a given infectious agent. Double stranded RNA which is produced during replication of several viruses is a potent inducer of IFN-I which then acts as a signal to induce expression of antiviral effectors, among which Adenosine Deaminase Acting on RNA (ADAR), a cellular protein which edits RNA [22]. RNA editing which does not include capping, 3' processing, and splicing is defined as post-transcriptional events that modify RNA molecules. It is a natural process used by mammal cells to generate different variants of a given protein. ADAR enzymes are capable of editing double stranded RNA ( [23] and for a recent review see [24]). There are three members of the mammalian ADAR family: ADAR1, ADAR2 and ADAR3. The formers convert adenosines into inosines, which are then translated into guanosines, while ADAR3 has no catalytic activity and seems mainly restricted to nervous system [25,26].
There are two ADAR1 isoforms. ADAR1-p110 is constitutively expressed and found exclusively in the nucleus, while ADAR1-p150 is an Interferon Stimulated Gene (ISG) induced by IFN-I [27]. ADAR1-p150 is found in the nucleus and in the cytoplasm [28,29]. These two isoforms are generated from two distinct promoters and an alternative splicing of the first exon. ADAR1-p110 and ADAR1-p150 have a C-terminal catalytic deaminase domain, three dsRNA binding domains and one Z-DNA binding domain in ADAR1-p110 or two in ADAR1-p150. Both ADAR1 isoforms edit cellular and viral RNAs possessing a doublestranded structure [30,31]. Interestingly, several viruses produce dsRNA during their replication cycle [22].
IFN-α is being used in combination to AZT to treat ATL patients, although the relationship between stimulation of innate immunity and impact on HTLV cycle are still not fully understood [21,54,55]. We recently demonstrated that PKR prevented HTLV-1 and -2 replication in cell culture following IFN-I treatment [11]. In another set of experiments using the exquisitely sensitive 3DI-PCR technique [56], we have also shown that HTLV-2 and simian T-cell leukemia virus type 3 (STLV-3) genomes can be edited by ADAR1 in cells transfected with HTLV-2 or STLV-3 molecular clones and an ADAR1-encoding plasmid [57].
However we could not find such mutations in ex vivo samples obtained from HTLV-2 infected individuals, thus demonstrating that ADAR1 editing of HTLV in vivo is a very rare phenomenon and is unlikely to play an antiviral role [57]. Though, consequences of ADAR1 expression on HTLV replication were not fully investigated in this report.
Here, we first demonstrate that ADAR1 is expressed in PHA/IL-2 activated primary T-lymphocytes in the absence of IFN-I stimulation, in HTLV-1 or HTLV-2 chronically infected T-cells and in primary CD4 + cells obtained from HTLV-1 infected individuals. We show that ADAR1 enhances HTLV-1 and HTLV-2 infection in T-lymphocytes and that this proviral effect is independent from its editing activity. ADAR1 expression suppresses IFN-α inhibitory effect on HTLV-1 and HTLV-2 through repression of PKR phosphorylation. Together with our previous results, these studies demonstrate that two interferon stimulated genes, i.e. PKR and ADAR1 have opposite effects on HTLV replication in vivo. The balanced expression of those proteins could determine the fate of viral cycle in the course of infection.

Results
ADAR1 is expressed in activated peripheral blood lymphocytes and in HTLV-1 and HTLV-2 chronically infected T cell lines ADAR1 was previously shown to edit the genome of some RNA viruses and to act therefore as an antiviral factor [31,34,44,45]. Consequently we first wondered whether ADAR1 was expressed in primary uninfected T-lymphocytes that are the target of HTLV-1 and -2 retroviruses in vivo. Given that HTLV-1 infected T-lymphocytes display an activated phenotype [58], we first obtained Peripherical Blood Lymphocytes (PBLs) from two healthy blood donors and activated them with PHA and IL-2. An aliquot of the same cells was also left quiescent. As expected, ADAR-p110 form was constitutively expressed, while ADAR1-p150 expression was induced following T-cell activation ( Figure 1A). As a control, primary B-cells were also purified from normal PBLs and left quiescent or activated with Pansorbin for 24 and 48 hours. As observed for T-cells, B-cell activation led to a strong increase in ADAR-1 expression ( Figure 1B), thus suggesting that increase in ADAR1 expression is linked to cell activation.
Then, we compared the expression of both ADAR1 isoforms in non-infected Jurkat T-lymphocytes as well as in HTLV-1 (C91-PL and MT2) or HTLV-2 (C19 and MO) infected T-lymphocytes ( Figure 1C). While ADAR1-p110 was expressed in all transformed cells including Jurkat, ADAR1-p150 was specifically up-regulated in HTLV-1 and HTLV-2 chronically infected cells.
Finally, CD4 + T-cells were isolated from the blood of two HTLV-1 infected individuals and stimulated with PHA and IL-2 for 3 days then grown only with IL-2. For each sample, ADAR-1 expression was then measured by flow cytometry in the fraction of CD4 + cells that express HTLV-1 (p19 gag ) antigen and compared to the p19 gag negative population. Interestingly, HTLV-1 productively infected cells expressed higher levels of ADAR1 than cells that were also stimulated but did not express HTLV-1 antigens ( Figure 1D).
Altogether, these results show that both isoforms of ADAR1 are expressed in the absence of IFN-I stimulation in activated T-cells that are susceptible to HTLV-1 and HTLV-2 infection. They also demonstrate that HTLV-1 and HTLV-2 chronically infected cells express the inducible ADAR1 isoform and suggest that HTLV-1 expression is correlated with an increased ADAR1 expression in primary CD4 + T-cells.
ADAR1 has a proviral effect on de novo HTLV-1 and HTLV-2 infection in T-lymphocytes Then, we investigated the effect of ADAR1 expression on de novo infection of T-cells with HTLV-1 and HTLV-2 using a previously described protocol (Figure 2A) [11].
HTLV-1/-2 infection occurs mainly through cell-cell contact. Target Jurkat ( Figure 2B, C) or primary PBLs ( Figure 2D, E) were therefore co-transfected with a plasmid encoding ADAR1 and a plasmid encoding the luciferase indicator gene under the control of the HTLV-1 or HTLV-2 LTR promoters. Twenty-four hours later, target cells were co-cocultured with gamma-irradiated C91-PL (HTLV-1), C19 (HTLV-2) donor cells or non-infected Jurkat cells. The viral non-structural Tax protein is necessary for activating transcription from the LTR. Consequently, luciferase activity monitors viral infection i.e. viral entry, reverse transcription, integration and viral expression in target cells (Figure 2A). ADAR1 expression in target Jurkat cells significantly enhanced luciferase activity after co-culture with HTLV-1 or HTLV-2 infected irradiated T-cells ( Figure 2B  Similar experiments were then performed using primary PBLs obtained from healthy blood donors as target cells, instead of Jurkat cells. As in Jurkat, ADAR1 expression in PBLs was associated with an increased luciferase activity (Figures 2D and E compare lane 2 vs. 4). To control that luciferase activity was linked to viral infection and not a passive Tax transfer from donor to target cells, we performed a series of control experiments where PBLs were treated or not with AZT, an inhibitor of reverse transcription. As expected, AZT treatment strongly reduced luciferase activity thus demonstrating that de novo Tax synthesis is required in the target cells ( Figure 2D and E, lanes 2 vs.3).
These results demonstrate that ADAR1 has a proviral effect on HTLV-1 and HTLV-2 de novo infection both in T-cell lines and in primary T-lymphocytes.

ADAR1 increases HTLV-1 and HTLV-2 protein expression
To decipher how ADAR1 affects HTLV-1/2 replication, Jurkat ( Figure 3A-B) or primary T-lymphocytes ( Figure 3C-D) were transfected with molecular clones which contain full length proviruses together with HTLV-1 or HTLV-2-LTR-luciferase reporter, in the presence or absence of ADAR1. This system allows us to bypass the entryintegration viral steps cells. In both settings, ADAR1 expression was associated with an increased luciferase activity, thus indicating up-regulation of Tax1 and Tax2 protein expression ( Figure 3A-D compare lane 2 and 3). Of note, in the presence of ADAR1 plasmid, a higher increase in luciferase activity was observed when Jurkat cells were transfected with the HTLV-1 molecular clone than with the HTLV-2 clone. This effect was not observed in primary Tcells.
293T-LTR-GFP cells were then transfected with HTLV-1 or HTLV-2 molecular clones. These cells stably harbor a GFP reporter gene whose expression parallels activation of LTR-driven transcription. Consistent with co-culture experiments, ADAR1 expression increased the number of GFP positive cells, thus demonstrating an increased Tax-mediated LTR transcription ( Figure 3E compare minus and plus ADAR1). ADAR1 expression was also associated with an increased number of syncitia, indicative of a higher HTLV envelope protein expression ( Figure 3E, compare minus and plus ADAR1). In addition, most GFP positive cells were also strongly positive for ADAR ( Figure 3F). Western-blot analyses, performed on 293T-LTR-GFP cells transfected with a constant amount of HTLV-1 ( Figure 3G) or HTLV-2 molecular clone ( Figure 3H) and increasing amounts of ADAR1 confirmed those observations. Indeed, both Tax1, Tax2 and p24 gag protein expression levels increased in a dose dependent manner ( Figure 3G-H). As a control, actin levels remain constant under the same conditions. The amount of p19 gag in the culture supernatant, which is assumed to reflect the quantity of viral particles, was also measured in the same experimental conditions. Consistent with western-blot results, p19 gag in the supernatant also increased in a dose-dependent manner ( Figure 3I, J).
Thus, these results demonstrate that ADAR1 increases viral expression and viral production in all cell lines tested.

ADAR1 expression does not impair the production of infectious viral particles
Given that ADAR1 possesses an editing activity, we then wondered whether viral particles produced in the presence of ADAR1 were infectious or not. 293T cells were transfected with HTLV-1 or HTLV-2 molecular clones and increasing amount of ADAR1 plasmid. Forty-eight hours after transfection culture medium was removed, cells were extensively washed before being co-cultured with Jurkat-LTR-luciferase cells ( Figure 4A). We observed a dose-dependent increase in luciferase activity, suggesting that viral particles that are made in the presence of ADAR1 are still capable of replication in their target cells ( Figure 4B, C). To control that luciferase activity measured in Jurkat target cells was not due to a passive Tax transfer from the transfected cell to the target cell, 293-T cells were transfected with Tax1 or Tax2 plasmids and co-cultured with Jurkat-LTRluciferase cells. Luciferase activity was measured and was similar to that of 293T cells transfected with a backbone vector ( Figure 4D, E).
Altogether, these results demonstrate that ADAR1 has a proviral effect, enhances HTLV-1/-2 expression and release of infectious particles.

ADAR1 proviral effect is independent from its editing activity
We next wondered whether ADAR1 proviral effect was dependent or not of its editing activity. Using a previously described ADAR1 RNA editing reporter system [40], we first verified that the plasmid encoding ADAR1 with a mutation in its catalytic site was deficient for its editing activity. The unedited reporter plasmid contains a stop codon that is converted into Trp codon when A to I editing by ADAR1 occurs. As a consequence, editing by ADAR1 is correlated with Firefly luciferase expression. As expected, transfection of wild-type ADAR1 induced an increased luciferase activity (Additional file 1: Figure S1 left graph), while ADAR1 mutant did not lead to any increase in luciferase expression (Additional file 1: Figure S1 right graph).
Then, 293-T cells were transfected with HTLV-1 or HTLV-2 molecular clones, in the presence of plasmids encoding either wild-type ADAR1 or the editing-deficient ADAR1 mutant. p24 gag western-blot ( Figure 5A, B) or p19 gag ELISA ( Figure 5C, D) revealed that expression of both wild-type or mutant ADAR1 led to a similar increase in viral protein expression and virus release.
Next, to confirm that editing has no effect on HTLV transcription, we used Jurkat cells transfected with the HTLV-LTR-luciferase reporter plasmids and co-transfected with HTLV-1 or HTLV-2 molecular clone, together with wild-type or editing-deficient ADAR1 constructs ( Figure 5E, F). A similar increase in luciferase activity was observed with of (B) pACH or (C) pH6neo molecular clones and increasing amount (0, 5, 50, 500 ng) of an ADAR1 expression plasmid or 8 μg of (D) pSG5M-Tax1, (E) pSG5M-Tax2 or backbone vector. Forty-eight hours later, cells were washed and co-cultured with Jurkat-LTR-luc cells for 24 hours (ratio 1:3). Luciferase activity results were normalized to protein concentration as determined by the Bradford method and calculated as fold change compared to cells non transfected with ADAR1 plasmid arbitrarily set to 100%. Data are the mean ± standard deviation (SD) from 3 independent experiments. Asterisks indicate statistically significant differences between treated and untreated cells (paired Student t test, *p < 0.05; ***p < 0.001; NS: non significant).
wild-type or mutant ADAR1. We also determined whether viral particles produced in the presence of both wild-type or mutant ADAR1 plasmids were infectious by performing a coculture assay as described in Figure 2A. ADAR1 mutant expression increased the release of infectious viral particles similarly to ADAR1 wild-type both in HTLV-1 and HTLV-2 settings (Additional file 2: Figure S2).
Altogether, these results confirm that ADAR1 proviral effect on HTLV-1 and HTLV-2 replication is independent from its editing activity.

ADAR1 proviral effect is mediated through inhibition of PKR activation
PKR is an ISG that has been shown to be involved, once activated by phosphorylation, in a signaling pathway leading to inhibition of mRNA translation. It was previously established that ADAR1 antagonizes PKR activation [37,49]. This effect requires double-strand RNA binding of ADAR [37]. We therefore wondered whether ADAR1 proviral effect on HTLV-1 and HTLV-2 could be mediated through PKR inhibition. To test this hypothesis, PKR western-blots were performed on protein extracts obtained from 293-T cells transfected with HTLV-1 and HTLV-2 molecular clones together with increasing amount of ADAR1 ( Figure 6A, B). Consistent with results presented above, the amount of p24 gag and Tax viral proteins increased in an ADAR1 dose dependent manner, while the level of PKR remained constant ( Figure 6A, B compare lane 1 vs 2-4). Interestingly, the level of phospho-PKR (i.e. the active form of PKR) was inversely correlated with that of ADAR1 ( Figure 6A, B, compare lane 1 and 4).
To demonstrate that ADAR1 proviral effect was linked to the inhibition of PKR activation, cells were transfected either with PKR siRNA or with control siRNA before being transfected with HTLV-1/2 molecular clones and ADAR1 plasmid ( Figure 6C, D). As previously described [11], PKR extinction results in an increased p24 expression. However, despite ADAR1 ectopic expression, the level of p24 gag remained constant, in cells transfected with PKR siRNA, while it increased when control siRNA were transfected.
These results demonstrate that ADAR1 effect on HTLV-1 and HTLV-2 is linked to the inhibition of PKR phosphorylation. A previous study demonstrated that HIV-1 Tat interacts with PKR and inhibited PKR auto-phosphorylation thus providing a possible mechanism of IFN suppression [59]. We therefore tested whether Tax could also bind and inhibit PKR. However no interaction between Tax and PKR was observed (data not shown).

ADAR1 expression reverses IFN-α inhibitory effect on HTLV-1 and HTLV-2
We have recently shown that IFN-α restricts HTLV-1 and HTLV-2 de novo infection by inhibiting viral protein synthesis through PKR activation [11]. Since ADAR1 inhibits PKR activation and enhances viral protein translation, we finally wondered whether IFN-α still had an antiviral effect in the presence of ADAR1 ectopic expression.
293-T cells were transfected with HTLV-1 ( Figure 7A) or HTLV-2 molecular clones ( Figure 7B), with or without an ADAR1 expression plasmid, before being treated with increasing amounts of IFN-α. In the absence of ADAR1 over-expression and consistent with our previous data [11], IFN-α treatment led to a decreased p24 gag as well as Tax1 or Tax2 protein expression ( Figure 7A and B, compare lane 1 and 4). This effect was correlated with an increase in PKR and phospho-PKR levels. In the presence of ectopic ADAR1, IFN-α treatment still lead to increased PKR levels, but the levels of phospho-PKR remained low and were not associated with a decreased p24 gag or Tax expression ( Figure 7A and B, compare lane 5 and 6-8).
These results show that IFN-α cannot activate PKR in the presence of elevated levels of ADAR1, thus ADAR1 expression prevents PKR from having its antiviral role.

Discussion
ADAR1 is responsible for A-to-G (I) editing of cellular and viral genes [30,31]. Editing can be site-selective in the case of Hepatitis virus D [41,42] or of L-glutamate receptor [60]. It can also be non-selective in the case of Hepatitis C, influenza, measles and Rift valley viruses [31,33,44]. As ADAR1 expression is induced by IFN-I, it was originally anticipated that IFN-I antiviral effects occur at least partly through an ADAR1 dependent editing of viral genomes. Indeed, Taylor et al. demonstrated that IFN-α treatment induces A-to-I editing in HCV replicon cells, while HCV editing is reduced when ADAR1 enzymatic activity is inhibited [33].
We previously demonstrated that co-transfecting ADAR1 together with HTLV-2 or STLV-3 molecular clones in 293-T cells led to A-to-G editing although the frequency of such editing could not be quantified [57]. A similar observation was made for measles and influenza viruses [44]. However, such hyper-edited sequences were only detected using the exquisitely sensitive 3DI-PCR technique [31], and replication of those viruses was not tested. In addition, editing was not observed in a series of samples obtained from 20 HTLV-2 infected individuals, although ADAR1 mRNA was present in those samples. This suggested that A-to-G (I) editing is an extremely rare phenomenon in vivo, at least in HTLV-2 infected cells [57].
As HTLV-1 mostly infects CD4 + T-lymphocytes and HTLV-2 infects CD8 + T-lymphocytes [16], we measured here ADAR1 p110 and p150 expression in peripheral blood lymphocytes that were left quiescent or PHA/IL-2 activated. Consistent with a previous study [61], both ADAR1 isoforms were present in activated lymphocytes. Thus, ADAR1 is expressed in cells that are targets of HTLV-1 and HTLV-2. Consistent with data obtained on HTLV-2 ex vivo samples [57], we have also shown that ADAR1 expression is higher in HTLV-1 and HTLV-2 chronically infected lymphocytes compared to noninfected T-cells and in primary CD4 + T-cells obtained from HTLV-1 infected individuals, thus suggesting that ADAR1 is expressed in response to HTLV infection. Interestingly, Clerzius et al have also shown that ADAR1-p110 and ADAR1-p150 expression is increased when Jurkat cells are infected with HIV-1 [50]. It could be of interest to determine in the future which factor induces ADAR1 expression in HTLV-1/-2 infected cells.
ADAR1 proviral effect was previously observed for different viruses [36,37,40,[49][50][51][52]. This effect is mediated through (1) viral RNA editing, (2) the inhibition of PKR through the formation of an ADAR1/PKR complex [37,50] or (3) an increase in ADAR1 catalytic activity. We have shown here that ADAR1 effect on HTLV-1/-2 replication is independent Figure 7 ADAR1 expression reverses IFN-α inhibitory effect on HTLV-1 and HTLV-2. 293-T cells were transfected with 4 μg of (A) pACH or (B) pH6neo molecular clone. When required, 500 ng of an ADAR1 expression plasmid were also transfected and cells were treated with increasing amount of IFN-α (0-1000 U/mL) for 48 hours. Western blot analyses were performed on 60 μg of proteins obtained from whole cell extracts using anti-ADAR1, anti-p24 gag , anti-Tax1, anti-Tax2, anti-PKR, anti-phospho-PKR and anti-actin antibodies. from ADAR1 editing activity and is rather linked to the inactivation of PKR function. It could also be of interest to determine whether ADAR1 effect is also linked to the formation of a complex between those two cellular proteins. Of note, our results do not exclude the fact that ADAR1 editing on HTLV-1/2 genomes might also occur, but at a very low frequency thus having no deleterious effect on viral cycle.
Shoggins et al have recently developed an assay to study the effect of hundreds of ISG products on different viruses that belong to Flaviviridae (YFV, WNV), Togaviridae (CHIKV, VEEV) or Retroviridae (HIV-1) [53]. This led to the identification of ISGs that are strong repressors, modest inhibitors, but also that are activators of viral cycle. In this study, ADAR was identified as an enhancer of replication for all viruses tested. Interestingly, the authors also tested the combination of one inhibitory and one enhancing ISG on viral cycle, and showed that, in most cases, viral replication was more influenced by the inhibitory than by the enhancing ISG [53]. Thus, this study highlighted that IFN-I does not necessarily promote an antiviral state and that a subtle balance between repressor and activator ISGs regulate the fate of viral infections.
We have previously shown that IFN-α inhibits HTLV-1 and HTLV-2 de novo infection through PKR activation [11]. We demonstrated here that another ISG, ADAR1 p150 is expressed in activated T-cells and plays a proviral role on HTLV-1 and HTLV-2 replication. We can therefore elaborate a working hypothesis that, during primary infection and in the absence of IFN-I, HTLV-1/-2 will infect cells that express ADAR1 p150 but not PKR. In this case, ADAR1 will play a proviral role. Infection could then lead to IFN-I production that could in turn activate both PKR and ADAR-1 p150 in surrounding cells. In this case, according to Schoggins et al., PKR could play an antiviral role [53]. Given that IFN-I is used to treat chronic ATL patients [62], it would therefore be interesting to measure quantitatively the expression of both PKR and ADAR in ATL cells if those samples were available.

Conclusions
In conclusion, our results demonstrate that IFN-I promotes the expression of two ISG that have antagonist roles on HTLV replication.

B-cells analyses
CD19 + B cells were purified by depletion/negative selection using a B-cell isolation kit (Miltenyi Biotech). Purity of isolated B-cells was monitored using anti-hCD19 VioBlue (Miltenyi Biotech) and was analyzed by fluorescence-activated cell sorting (FACS CantoII; BD Biosciences).

T-cells analyses
CD4 + T-cells obtained from healthy donors or HTLV-1 infected individuals were purified by depletion/negative selection using T-cell isolation kit (Miltenyi Biotech). Purity of CD4 + T-cells was monitored using anti-hCD4APC antibodies (Miltenyi Biotech) and was analyzed by fluorescence-activated cell sorting (FACS CantoII; BD Biosciences). CD4 + T-lymphocytes were then cultured at 1.10 6 /ml with IL-2 (150 U/ml) and PHA (1 μg/ml) for 3 days and then maintained only with IL-2 for 5 days. Cells were analyzed with a surface staining against hCD4-APC and hCD69-PE (Miltenyi Biotech) or intracellular staining for ADAR and p19 (Zeptometrix) after permeabilization/ fixation with the Cytofix/Cytoperm kit (BD). Cells were analyzed on a FACS-Canto II (BD Sciences) collecting 100.000 events. Results were analyzed using the FlowJo software.

Fluorescence microscopy
293T-LTR-GFP cells were seeded at 5. 10 5 cells/well on 6well plates. The next day, one microgram of the appropriate plasmid was transfected as described above, with or without the ADAR plasmid. Two days later, cells were visualized and pictures were captured with a Zeiss Axio Vert.A1 imaging microscope.

p19 gag ELISA
Cell culture supernatants were collected 48 hours posttransfection and the presence of the HTLV matrix p19 gag protein was quantified using a commercially available ELISA kit (Retrotek HTLV-1/2 p19 antigen, Zeptometrix) following the manufacturer's instructions as previously described [11]. In brief, wells are coated with polyclonal antibodies which react with the major gag gene products of HTLV-1 and HTLV-2. Viral antigen in the test specimen is captured by the antibody.

PKR silencing
293T cells were seeded at a concentration of 3.10 5 cells per well onto 6-well plates. The following day, 20 nM of PKR siRNA (ON-TARGETplus SMART pool EIF2AK2, Fermentas) or control siRNA (ON-TARGETplus Non-targeting Pool, Fermentas) were transfected (HiPerfect reagent, Qiagen) as previously described [11]. Twelve hours posttransfection, 1.2 μg of pACH or PH6neo plasmid were transfected together with 20 nM of siRNA and 125 ng of ADAR1 (Attracten reagent, Qiagen) following the manufacturer's instructions. Cells were collected 48 hours later and analyzed.

Additional files
Additional file 1: Figure S1. Testing the editing capacity of ADAR1wt and ADAR1 mutant constructs. 293-T cells were transfected with 150 ng of the 3XF-ADAR1 RNA editing reporter system with (left) 0, 150, 300, 600, 1000 ng of the plasmid coding for ADAR1 wild-type or with (right) 0, 150, 300, 600, 1000 ng of the plasmid encoding ADAR1 with a mutation in its catalytic site. Forty-eight hours later, luciferase activity was measured and results normalized to renilla expression and calculated as fold change compared to cells non transfected with ADAR1 plasmid arbitrarily set to 100%. Data are the mean ± standard deviation (SD) from 2 independent experiments. Asterisks indicate statistically significant differences between treated and untreated cells (paired Student t test, *p < 0.05; NS: non significant).
Additional file 2: Figure S2. Viral particles produced in the presence of an editing-deficient ADAR1 are still infectious. 293-T cells were transfected with 4 μg of (A) ACH or (B) pH6neo molecular clone together with 500 ng of wild type or mutant ADAR1. Forty-eight hours later, cells were washed and co-cultured with Jurkat-LTR-luc cells for 24 hours (ratio 1:3). Luciferase activity was normalized by protein concentration as determined by the Bradford method and calculated as fold change compared to cells non transfected with ADAR1 plasmid arbitrarily set to 100%. Paired Student t test, NS: not significant.