Although it is well established that the release of HCV (hepatitis C virus) occurs through the secretory pathway, many aspects concerning the control of this process are not yet fully understood. α-Taxilin was identified as a novel binding partner of syntaxin-4 and of other members of the syntaxin family, which are part of SNARE (soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor) complexes and so are involved in intracellular vesicle traffic. Since α-taxilin prevents t-SNARE (target SNARE) formation by binding exclusively to free syntaxin-4, it exerts an inhibitory effect on the vesicular transport. HCV-replicating Huh7.5 cells and HCV-infected primary human hepatocytes and liver samples of patients suffering from chronic HCV contain significantly less α-taxilin compared with the controls. HCV impairs the expression of α-taxilin via NS5A-dependent interruption of the Raf/MEK [MAPK (mitogen-activated protein kinase)/ERK (extracellular-signal-regulated kinase) kinase] signal transduction cascade. Moreover, the half-life of α-taxilin is significantly reduced in HCV-replicating cells. Whereas modulation of α-taxilin expression does not significantly affect genome replication, the overexpression of α-taxilin prevents the release of HCV. In contrast with this, silencing of α-taxilin expression leads to increased release of infectious viral particles. This is due to the negative effect of α-taxilin on t-SNARE formation that leads to impaired vesicular trafficking. Accordingly, overexpression of the t-SNARE component syntaxin-4 increases release of HCV, whereas silencing leads to an impaired release. These data identify α-taxilin as a novel factor that controls the release of HCV and reveal the mechanism by which HCV controls the activity of α-taxilin.
HCV (hepatitis C virus) is a major cause of chronic hepatitis, liver cirrhosis and hepatocellular carcinoma. Worldwide, there are more than 170 million people chronically infected with HCV. HCV is a member of the Flaviviridae family that contains a single-stranded positive-sense RNA genome of approximately 9600 bases [1,2].
The viral genome is translated into a polyprotein of approximately 3100 amino acids which is co- and/or post-translationally processed by cellular and viral proteases. The N-terminus encompasses the structural proteins core, E1, and E2, the C-terminus, the p7 protein and the NS (non-structural) proteins NS2, NS3, NS4A, NS4B, NS5A and NS5B [3–5].
HCV replication takes place at specialized rearranged intracellular membranes derived from the ER (endoplasmic reticulum), the so-called ‘membranous web’ [6–8]. Proteins involved in VLDL (very-low-density lipoprotein) assembly, i.e. apoB (apolipoprotein B), apoE (apolipoprotein E) and MTP (microsomal triacylglycerol-transfer protein)  are enriched in the membranous web. Previously, it was observed that NS5A interacts directly with the sorting factor TIP47 (47 kDa tail-interacting protein) [10,11] and apoE [12–14].
LDs (lipid droplets), intracellular lipid-storing organelles, play a crucial role for the HCV morphogenesis [3,6,15–17]. In HCV-replicating cells, the surface of LDs is coated with the viral core protein. Core proteins are targeted to the LDs via DGAT1 (diacylglycerol acyltransferase 1) [18,19] and recruit NS proteins to LD-associated membranes, a crucial step for virus morphogenesis. It was found that apoE and TIP47 bind directly to NS5A and are essential for replication, virus morphogenesis and release of infectious viral particles [10,11,13,20]. Although it is well established that the release of HCV occurs through the secretory pathway, the control of this process has not yet been fully understood [12,21–23].
α-Taxilin was identified as a novel binding partner of the syntaxin family, which is involved in intracellular vesicle trafficking. Syntaxin-1a, syntaxin-3 and, especially, syntaxin-4 bind to α-taxilin. However, α-taxilin exclusively interacts with free syntaxins that are not part of a SNARE (soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor) complex . SNARE proteins are a large protein superfamily consisting of more than 60 members in yeast and mammalian cells. SNARE complexes in human cells are formed by nine different proteins encompassing synaptobrevins and syntaxins.
SNARE proteins are involved in the intracellular trafficking of vesicles mediating the fusion of vesicles with their target membrane. Two categories of SNAREs can be distinguished. On the one hand v-SNAREs (vesicle SNAREs) that are found on the membrane of the transport vesicles and on the other hand t-SNAREs (target SNAREs) that are localized on the membrane of the target compartment. The SNARE motif that consists of 60–70 amino acids and contains heptad repeats is common to all SNAREs and confers the ability to form coiled-coil structures. Reversible formation of the trans-SNARE complexes by v- and t-SNAREs is the crucial step for SNARE-dependent membrane fusion [25,26].
The α-taxilin–syntaxin interaction can be blocked by SNAP-25 (25 kDa synaptosome-associated protein) or Munc18 . On this basis, it is assumed that α-taxilin acts as a negative regulator of vesicular transport by withdrawing syntaxin-4 and thereby impairing the assembly of t-SNARE complexes. α-Taxilin encompasses 546 amino acids and has a calculated molecular mass of 61.9 kDa . Previously, it was observed that the expression of α-taxilin is strongly induced in HBV (hepatitis B virus)-replicating cells . α-Taxilin is essential for the release of HBV by mediating the interaction between the HBV surface protein LHB (large HBV envelope polypeptide) and the ESCRT (endosomal sorting complex required for transport) component tsg101 (tumour susceptibility gene 101). Silencing of α-taxilin expression abolishes the release of viral particles that occurs via MVBs (multivesicular bodies), while the release of subviral particles that occurs via the classic secretory pathway is increased .
MATERIALS AND METHODS
The Huh7-derived cell clone Huh7.5 , which is highly permissive for HCV RNA replication, was used for transfection and infection assays as described in . Isolation of primary human hepatocytes, cultivation and infection were performed as described in [31,32]. Huh7 cell clone 9-13 harbouring the HCV replicon I377/NS3-3′ has been described previously . The human hepatocyte cell lines HepG2 and Huh7 were used to study the release of VLDLs and the effect of IFN-α (interferon α) respectively.
The following commercial primary antibodies were used: anti-NS3 (8G-2, Abcam), anti-β-actin for detection of human and mouse protein (AC-74, Sigma–Aldrich), anti-core (MA1-080, Thermo Scientific), anti-α-taxilin (H-66 and E-2, Santa Cruz Biotechnology), anti-p62 (GP62-C, Progen Biotechnik) and anti-syntaxin-4 (ab57841, Abcam). For detection of NS5A, a polyclonal rabbit derived serum was used . Alexa Fluor® 488/546-conjugated secondary antibodies were obtained from Invitrogen. Cy3 (indocarbocyanine)-conjugated antibodies were purchased from Jackson ImmunoResearch Laboratories, secondary antibodies for Western blotting were obtained from LI-COR and horseradish peroxidase-conjugated secondary antibodies were purchased from GE Healthcare.
Plasmids pFK-JHF1/wt, pFK-JHF1/J6 and pFK-JHF1/GND have been described in . pDEST26-TXLNA was purchased from ImaGenes. Luciferase reporter construct harbouring the promoter of α-taxilin is described in . For analysis of ISRE (IFN-stimulated-response element)-dependent gene expression, pISRE-luc (Stratagene) was used as reporter construct. Plasmids coding for constitutive active Raf (pv-Raf) and for a trans-dominant-negative Raf mutant (ptdnRaf) were described in . The plasmid pHBx coding for the HBx protein has been described in . Double-stranded oligonucleotides were synthesized using the primers 5′-CACCG-CCACCATGGGGTACCCATACGACGTCCCAGACTACGCT-CGGGACAGGACCCACGAG-3′ and 5′-TTATCCAACCAC-TGTGACGC-3′ and the cDNA of Huh7.5 cells as template. This sequence was inserted into the vector TOPO-pCDNA3.1 (Invitrogen) to get the construct pCDNA3.1-STX4 coding for HA (haemagglutinin)-tagged syntaxin-4. For HBV expression, a 1.2-fold HBV-genome ayw (pHBV1.2) was used. The plasmid pCDNA3.1-NS5A coding for the NS5A protein has been described in .
Real-time PCR (rtPCR)
RNA isolation was performed using peqGOLD TriFast (PEQLAB Biotechnologie) according to the manufacturer's instructions. cDNA synthesis and rtPCR were performed as described recently . For primers, see Supplementary Table S1.
Quantification of viral genomes in the cell culture supernatant
Viral genomes in the cell culture supernatant were quantified as described in .
In vitro transcription and RNA transfection
In vitro transcription, electroporation of HCV RNAs, and luciferase assays were performed as described in . All luciferase assays were done at least in triplicate.
The cells were transfected 24 h after electroporation using linear PEI (polyethyleneimine) (Polysciences) as described in  and incubated for a further 48 h or 72 h.
For transient RNAi transfection, cells were transfected with 20 nM α-taxilin-specific (sc-39644) siRNA or 20 nM syntaxin-4-specific (sc-36590) siRNA (both from Santa Cruz Biotechnology) using a N-TER™ Nanoparticle siRNA Transfection System (Sigma) according to the manufacturer's protocol. Cells were analysed 96 h post-transfection.
Determination of half-life
Half-lives were determined as described in . Huh7.5 cells were transfected with JFH1/GND or JFH1/J6 and cultured for 14 days to allow high reinfection rates. Cells were treated with 35.5 μM cycloheximide to inhibit protein synthesis and lysed after 1, 2, 3 and 4 h of treatment. The amount of α-taxilin and β-actin in cell lysates was determined by Western blotting and densitometric scans. The relative amount of α-taxilin was normalized to the untreated control and the half-life was estimated by exponential regression. Six independent experiments were performed.
Cells were treated with 100 nM bafilomycin A1 (Sigma) for 24 h to inhibit autophagy.
Cells were stimulated for 8 h with 1000 units of human IFN-α B2 (Tebu-Bio) before RNA isolation or luciferase activity assay.
Indirect immunofluorescence analysis
Analysis of indirect immunofluorescence was performed as described in . Immunofluorescence staining was analysed using a confocal laser-scanning microscope (CLSM 510 Meta, Carl Zeiss) and ZEN 2009 software.
The tissue samples were taken for routine diagnostics during surgery. Tissue which was not needed any more for diagnostic approaches was used for staining in this study. The patients gave informed consent for histological and immunohistological examinations. Three uninfected and three HCV-positive tissue samples were analysed. The fluorescence intensity was analysed with the software ZEN2009 (Zeiss). Laser intensity and digital gain were adjusted to one reference sample and kept constant in all samples.
The total amounts of cholesterol, HDL, LDL and triacylglycerols in cell culture supernatant were measured in the University Hospital Frankfurt, Frankfurt am Main, Germany. The amount of VLDL was calculated by the Friedewald equation .
Results are described as means±S.E.M. from at least three experiments. The significance of results was analysed by one-tailed ratio Student's t test using GraphPad Prism version 5.04 for Windows. Error bars are S.D. *P<0.05, **P<0.01, ***P<0.001. In histograms showing relative changes compared with the control, the control group was arbitrarily set as 1. Here, an S.D. for the control group cannot be reported, as standardization of the measured values (relative to the control group) was performed for each of the independent assays. Therefore measurements for the treatment groups in each assay were dependent (matched).
Decreased amount of α-taxilin in HCV-replicating cells
Gene expression profiling of the liver of NS5A transgenic (NS5Atg) mice identified α-taxilin as one of the strongest repressed genes in these mice . Western blot analyses of liver lysates derived from these mice revealed a significantly decreased amount of α-taxilin compared with the wild-type animals (Figure 1A). To investigate whether this can be observed for HCV-replicating cells as well, HCV-positive cells were generated by transfection of HuH7.5 cells with the replication-competent HCV genomes JFH-1 or J6. The replication-deficient genome GND served as control. Western blot analysis of cellular lysates derived from these cells using an α-taxilin-specific antiserum revealed a significantly decreased amount of α-taxilin in HCV-replicating cells compared with the control (Figure 1B). This could be confirmed by immunofluorescence microscopy of HCV-replicating Huh7.5 cells: HCV-positive cells were visualized by core-specific staining (green) and possess a significant smaller amount of α-taxilin (red) compared with the HCV-negative control cells (Figure 1C). Immunofluorescence microscopy of liver samples derived from three different patients suffering from chronic HCV infection confirmed the decreased α-taxilin amount by comparison with liver samples from three HCV-negative patients (Figure 1D). The quantification shows a significantly (P ≤ 0.01) lower amount of α-taxilin in the liver tissue of HCV-positive patients compared with the HCV-negative patients (Figure 1D).
Decreased amount of α-taxilin in HCV-replicating cells
Quantification of α-taxilin-specific transcripts in HCV-replicating primary human hepatocytes and the corresponding controls by rtPCR revealed that the expression of α-taxilin is moderately repressed in these cells (Figure 1E).
These data indicate that in HCV-replicating cells, the amount of α-taxilin is decreased.
HCV represses the expression of α-taxilin and destabilizes α-taxilin
To study the mechanism leading to a decreased amount of α-taxilin in HCV-replicating cells, α-taxilin-specific transcripts were quantified by rtPCR. Compared with the control, the rtPCR revealed a moderately decreased number of α-taxilin-specific transcripts in HCV-replicating cells (Figure 2A). For further analyses, a reporter gene construct harbouring the luciferase reporter gene under control of the α-taxilin promoter was used (ppromataxluc ). Transfection of HCV-replicating cells (JFH1 and J6) with ppromataxluc revealed a significantly decreased luciferase activity compared with HCV-negative cells (GND) (Figure 2B). Previous data indicated that activation of the taxilin promoter can be triggered by c-Raf-dependent signalling cascades . Co-transfection of a constitutively active Raf mutant (v-Raf) or with the established activators of the c-Raf signalling cascade such as HBV (pHBV1.2 harbours a 1.2-fold HBV genome) or with the HBV regulatory protein HBx (pHBx) with the reporter construct triggers a strong activation of the reporter construct (Figure 2C). However, co-expression of a tdn mutant of c-Raf (Raf C4) abolishes the HBx-dependent activation of the taxilin promoter, demonstrating the relevance of functional c-Raf for activation of the α-taxilin promoter (Figure 2C). NS5A was reported to interrupt the c-Raf-dependent signalling cascade . Accordingly, co-transfection of Huh7.5 cells with ppromataxluc and with a NS5A expression vector (pcDNA3.1-NS5A ) leads to decreased luciferase activity compared with the control-transfected cells (Figure 2D).
HCV impairs expression of α-taxilin and shortens the half-life of α-taxilin
These data indicate that HCV decreases the expression of α-taxilin via NS5A-dependent interruption of c-Raf signalling.
Since the reduction of the α-taxilin expression in HCV-positive cells is moderate, we asked whether further mechanisms contribute to the decrease of the amount of α-taxilin in these cells.
Interestingly, the analysis of half-life (t½) of α-taxilin revealed that the stability of α-taxilin is significantly decreased in HCV-replicating cells compared with HCV-negative control cells. The t½ in HCV-replicating cells was determined at 3.2 h compared with 6.8 h in HCV-negative Huh7.5 cells (Figure 2E).
Taken together, these data indicate that the HCV-dependent reduction in the amount of α-taxilin is due to an NS5A-mediated impaired expression and due to a decreased half-life.
Overexpression of α-taxilin impairs release of infectious particles from HCV-replicating cells
The strong deregulation of the amount of α-taxilin in HCV-replicating cells suggests a relevance of α-taxilin for the viral life cycle. Since α-taxilin was described to interfere with the formation of t-SNARE complexes, it was investigated whether overexpression of α-taxilin affects the release of infectious viral particles. HCV-replicating cells (J6) were transfected with an expression vector encoding α-taxilin (pDEST26-TXLNA) or the corresponding empty control vector. The amount of released infectious viral particles was quantified by determination of the TCID50 of the supernatant. The TCID50 value shows that overexpression of α-taxilin indeed leads to a 10-fold reduction of the amount of released infectious viral particles (Figure 3A).
Overexpression of α-taxilin impairs the release of infectious HCV
To investigate whether this is due to an impaired release or due to a reduced replication, the intracellular amount of viral genomes and of NS protein (NS3) and structural protein (core) was determined (Figures 3B and 3C). rtPCR revealed for the viral genomes a slight (2-fold) reduction in the case of α-taxilin-overexpressing cells compared with the control (Figure 3B). The quantification of the Western blots shows that overexpression of α-taxilin in HCV-replicating (J6) cells leads to a slight reduction in the intracellular amount of core and NS3 (Figure 3C).
Since HCV is known to induce autophagy [47,48], it was investigated whether the autophagosomal activity is involved in the reduction of the intracellular amount of NS3 or core in α-taxilin-overexpressing cells. To investigate this, autophagy was inhibited by bafilomycin. Indeed, quantification of the Western blot analysis of lysates derived from bafilomycin-treated cells and untreated control cells shows that the inhibition of autophagy by bafilomycin restores the intracellular amount of core and NS3 in taxilin-overexpressing cells up to the value observed for the control-transfected cells (Figure 3D). This suggests that overexpression of α-taxilin primarily impairs the release of de novo synthesized viral particles and less so the replication.
To address this point in more detail, the Huh7 cell clone 9-13 harbouring the subgenomic replicon I377/NS3-3′ was used. This subgenomic replicon expresses the NS proteins and thereby is able to perform genome replication, but lacks the structural proteins and so fails to assemble/release viral particles. This allows analysis of the effect of α-taxilin on replication in the absence of any interference with viral morphogenesis/release. The Western blot analysis of cellular lysates derived from cells replicating the subgenomic replicon shows that overexpression of α-taxilin does not significantly affect the amount of NS3 and NS5A (Figure 3E) compared with the control. Moreover, there was no significant effect on the replication as determined by quantification of the subgenomic RNA (Figure 3F). Taken together, these data indicate that overexpression of α-taxilin primarily impairs the release of infectious viral particles.
Interferon α does not induce the expression of α-taxilin
In the light of the inhibitory effect of α-taxilin on the release of infectious HCV particles, we asked whether α-taxilin can be considered as an IFN-dependent restriction factor. To investigate this, the effect of IFN-α on the expression of α-taxilin was studied. Analysis of α-taxilin expression by rtPCR of IFN-α-treated Huh7.5 cells revealed no elevated amount of α-taxilin-specific transcripts. Quantification of MxA-specific transcripts served as positive control. The expression analysis revealed that no IFN-α-dependent induction of the α-taxilin expression could be observed (Figure 3G). To exclude that this is due to a block in IFN-dependent signalling in Huh7.5 cells, the effect of IFN-α stimulation on α-taxilin expression was analysed by reporter gene assay in Huh7 cells. Huh7 cells were transfected with the reporter construct ppromataxluc or as control with the pISRE-luc vector and treated with IFN-α. The reporter gene assay revealed that, compared with the ISRE control, the taxilin promoter is not stimulated by IFN-α (Figure 3H). This was confirmed by rtPCR for quantification of α-taxilin- and MxA-specific transcripts. The rtPCR confirmed that the expression of α-taxilin in Huh7 cells is not induced by IFN-α (Figure 3I). Taken together, these data indicate that the expression of α-taxilin is not IFN-α-dependently regulated.
Silencing of α-taxilin expression increases release of infectious viral particles from HCV-replicating cells
The data described above indicate an inhibitory effect of α-taxilin on the release of viral particles. To test the hypothesis that impaired expression of α-taxilin favours the release of HCV particles, the amount of α-taxilin was decreased by α-taxilin-specific siRNA. The amount of released infectious viral particles was quantified by determination of the TCID50 of the supernatant. The TCID50 value shows that silencing of the α-taxilin expression indeed leads to a significant increase in the amount of released infectious viral particles (Figure 4A).
Silencing of α-taxilin expression increases the amount of released infectious viral particles
To investigate whether this is primarily due to an effect on the release or due to an elevated replication, the intracellular amount of NS protein (NS3) and structural protein (core) was determined (Figure 4B). Silencing of α-taxilin expression leads to a slightly increased intracellular amount of NS3 and core (Figure 4B). To control the specificity of the observed effect, expression of α-taxilin was rescued by co-transfection with the expression vector pDEST26-TXLNA. Rescue of α-taxilin expression by co-transfection with an α-taxilin expression vector abolished the increased intracellular levels of core and NS3 (Figure 4C).
To study the direct effect of impaired α-taxilin expression on genome replication in the absence of virus morphogenesis/release, α-taxilin expression was silenced in cells replicating a subgenomic replicon (Figures 4D and 4E). Although the amount of NS3 was not significantly affected by silencing the α-taxilin expression compared with the control, the amount of NS5A was slightly increased (Figure 4D). Analysis of the intracellular amount of subgenomic genomes revealed that silencing of α-taxilin expression led to a slight increase in the amount of subgenomic RNAs compared with the control (scrRNA) (Figure 4E).
Taken together, these data demonstrate that the release of infectious viral particles from HCV-replicating cells is increased if the level of α-taxilin is decreased.
Modulation of syntaxin-4 expression affects HCV release
The data described above demonstrate that in HCV-replicating cells, the amount of α-taxilin is decreased. α-Taxilin was identified as binding partner of free syntaxin-4. Since α-taxilin binds exclusively to free syntaxin-4 that is not part of the SNARE complex, α-taxilin prevents the formation of t-SNARE complexes. In the light of this, we hypothesize that HCV decreases the amount of α-taxilin to favour t-SNARE complex formation, thereby facilitating vesicular trafficking required for the release of viral particles.
To test this hypothesis, HCV-replicating cells were transfected with the expression vector pCDNA3.1-STX4. Western blot analyses of cellular lysates revealed a reduced intracellular amount of core while the level of NS3 was almost unaffected (Figure 5A). Quantification of the amount of viral RNA in the supernatant by rtPCR demonstrated that overexpression of syntaxin-4 increases the amount of released viral genomes (Figure 5B).
Modulation of syntaxin-4 expression in HCV-replicating cells influences the amount of released viral particles
Conversely, the syntaxin-4 expression was impaired by specific siRNA. Western blot analyses of cellular lysates revealed an increased intracellular amount of core while the level of NS3 was almost unaffected (Figure 5C). Quantification of released viral genomes revealed that inhibition of syntaxin-4 expression decreases the amount of released genomes (Figure 5D).
To investigate further the interplay between α-taxilin and syntaxin-4 with respect to the HCV life cycle, HCV-replicating cells were co-transfected with a syntaxin-4 expression vector and an α-taxilin expression vector. No significant change in the amount of viral particles released was observed since the increased amount of syntaxin-4 is neutralized by the increased amount of α-taxilin that acts as an inhibitor by binding to the overexpressed syntaxin-4 (Figures 5E and 5F).
Conversely, if the expression of syntaxin-4 and α-taxilin was simultaneously impaired by specific siRNAs, the release of infectious viral particles was decreased. The stimulatory effect of a decreased amount of α-taxilin was overcompensated for by the decreased amount of syntaxin-4 (Figures 5G and 5H). Under these conditions, syntaxin-4 as an effector is unable to form functional SNARE complexes that are required for the release of infectious HCV particles.
Taken together, these data indicate that α-taxilin is a novel factor controlling the release of HCV particles by affecting the level of free syntaxin-4 and thereby controlling t-SNARE complex formation.
The release of HCV is tightly associated with the VLDL secretion machinery. On the basis of RNAi screens, VAMP1 (vesicle-associated membrane protein 1) was identified as an essential factor for release of infectious HCV particles . VAMP1 is found on secretory vesicles as a v-SNARE component which mediates the fusion with the target plasma membrane by interaction with syntaxin-4 as a component of the t-SNAREs.
In the present study, α-taxilin was identified as a novel factor that affects the release of HCV from HCV-replicating cells. In previous papers, α-taxilin was described to bind to members of the syntaxin family, especially syntaxin-4 [24,27]. Syntaxin-4 is part of the t-SNARE complex. α-Taxilin was described to exclusively bind to free syntaxin-4 that is not part of the SNARE complex . On this basis, it is assumed that α-taxilin exerts an inhibitory effect on vesicular transport processes that depend on SNARE-mediated fusion: binding of α-taxilin to syntaxin-4 reduces the amount of free syntaxin-4 and thereby prevents formation of SNARE complexes. Indeed, the release of HBV subviral particles (HBsAg) that occurs via the ER–Golgi secretory pathway  is impaired if α-taxilin is overproduced, whereas silencing of the expression of α-taxilin leads to an increased release of HBsAg . In contrast, release of hepatitis B viral particles that occurs ESCRT-dependently via MVBs is increased when α-taxilin is overproduced .
The reduction of the amount of α-taxilin or overexpression of syntaxin-4 facilitates t-SNARE complex formation and vesicular trafficking and thereby the release of HCV.
In accordance to this observation, HCV decreases the amount of intracellular α-taxilin by several mechanisms. Previously, it was found that the expression of α-taxilin can be modulated by activation of the c-Raf/MEK/ERK [where ERK is extracellular-signal-regulated kinase and MEK is MAPK (mitogen-activated protein kinase)/ERK kinase] signalling cascade . The HCV protein NS5A that interrupts MEK/ERK activation [31,34,50] is able to impair the expression of α-taxilin. This could be observed in HCV-replicating cells as well as in NS5A-overexpressing cells and in the liver of NS5Atg mice . Apart from the reduced expression of α-taxilin, the level of α-taxilin is diminished further by a significantly shorter half-life of α-taxilin in HCV-replicating cells. The relevance of a reduction of the free α-taxilin level for the release of infectious viral particles was demonstrated by modulation of α-taxilin expression. Overexpression of α-taxilin leads to an impaired release, whereas silencing of α-taxilin expression increases the amount of released viral particles. Conversely, overexpression of syntaxin-4 leads to an increased secretion of HCV particles, whereas silencing of syntaxin-4 leads to a decreased release of infectious viral particles. This reflects the interplay between syntaxin-4 as a t-SNARE component and α-taxilin controlling the release of HCV particles.
Modulation of the amount of α-taxilin by overexpression or silencing of the expression does not significantly affect genome replication. This was corroborated by studies based on subgenomic replicon systems that allow the study of the interaction with HCV replication in the absence of any interference with virus morphogenesis or release. On the basis of these observations, we conclude that the major mode of action of α-taxilin is the inhibition of the particle release.
In the light of the inhibitory effect of α-taxilin overexpression on the release of infectious HCV particles, one might ask why overexpression of α-taxilin does not lead to a significant intracellular accumulation of HCV. It was described that only a small amount of the de novo synthesized viral particles are finally released as infectious viral particles. The large majority of the de novo synthesized viral particles is prevented from release and intracellularly degraded by the autophagosomal route . Indeed, inhibition of autophagy leads to a significant increase in the intracellular amount of core in α-taxilin-overexpressing cells compared with the control-transfected cells. This indicates that α-taxilin interferes with the release process at a late stage, affecting only the small fraction of release-prone infectious viral particles.
In the light of the inhibitory effect of α-taxilin on the release of HCV particles from infected cells, it was tempting to speculate whether α-taxilin could act as an IFN-dependently expressed restriction factor. However, stimulation of Huh7.5 or HepG2 cells with IFN-α provided no evidence for an IFN-dependent induction of α-taxilin expression.
Although overexpression of α-taxilin or decrease of syntaxin-4 expression leads to a decreased amount of released infectious viral particles, no intracellular accumulation could be observed by immunofluorescence microscopy. This might again reflect that only a small fraction of the de novo synthesized viral proteins finally become part of released infectious viral particles . In the light of this, it could be hypothesized that a late step in the release process is controlled by α-taxilin, affecting exclusively the release of the small fraction of properly assembled release-prone infectious viral particles.
Release of HCV is tightly associated with the release of VLDL. Overexpression of α-taxilin in HepG2 cells is associated with a decreased release of VLDL (results not shown). The relevance of α-taxilin in controlling, e.g., VLDL release by negative regulation of SNARE formation argues for the hypothesis that HCV particles are released by the classic release pathway.
Taken together, these data indicate that α-taxilin is a novel crucial factor controlling the release of HCV particles. In the light of its strong inhibitory potential on the release of HCV particles, the virus has evolved a variety of strategies to decrease the intracellular amount of free α-taxilin.
Christian Donnerhak, Daniela Ploen, Kiyoshi Himmelsbach, Fabian Elgner and Eberhard Hildt designed the experiments. Christian Donnerhak, Fabian Elgner, Huimei Ren, Regina Medvedev, André Schreiber, Lorenz Weber, Markus Heilmann, Daniela Ploen, Kiyoshi Himmelsbach and Eberhard Hildt performed experiments. Christian Donnerhak, Fabian Elgner and Eberhard Hildt wrote the paper. Isolation of primary human hepatocytes was performed by Malin Finkernagel.
We thank Andrea Henkes and Gert Carra for their excellent technical support and Dagmar Fecht-Schwarz for critically reading the paper before submission.
This work was supported by a grant from the German Center for Infection Research (DZIF) to E.H.
hepatitis B virus
hepatitis C virus
MAPK (mitogen-activated protein kinase)/ERK kinase
soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor
median tissue culture infective dose
47 kDa tail-interacting protein
vesicle-associated membrane protein
These authors contributed equally to this work.