2 the establishment of replication complexes. 4 5 6 7 8 9...

1

Serine phosphorylation of the hepatitis C virus NS5A protein controls 1

the establishment of replication complexes. 2

3

Douglas Ross-Thriepland, Jamel Mankouri and Mark Harris * 4

School of Molecular and Cellular Biology, Faculty of Biological Sciences and Astbury 5

Centre for Structural Molecular Biology, University of Leeds, Leeds, LS2 9JT, UK. 6

* Address for correspondence: School of Molecular and Cellular Biology, Faculty of 7

Biological Sciences, University of Leeds, Leeds LS2 9JT, UK. Tel +44 113 343 5632, 8

e-mail: [email protected] 9

Running title: Phenotype of NS5A serine phosphorylation 10

Word count: Abstract 225 (importance: 82), main text (including figure legends): 6951 11

8 Figures, 2 supplementary figures, 3 supplementary videos 12

13

JVI Accepts, published online ahead of print on 31 December 2014J. Virol. doi:10.1128/JVI.02995-14Copyright © 2014, American Society for Microbiology. All Rights Reserved.

on June 24, 2018 by guesthttp://jvi.asm

.org/D

ownloaded from

http://jvi.asm.org/

2

ABSTRACT. The hepatitis C virus (HCV) non-structural 5A protein (NS5A) is highly 14

phosphorylated and involved in both virus genome replication and virion assembly. We, and 15

others, have identified serine 225 in NS5A as a phosphorylation site but the function of this 16

post-translational modification in the virus lifecycle remains obscure. Here we describe the 17

phenotype of mutations at serine 225, this residue was mutated to either alanine (S225A: 18

phosphoablatant) or aspartic acid (S225D: phosphomimetic) in the context of both the JFH-1 19

cell culture infectious virus and a corresponding subgenomic replicon. S225A exhibited a 10-20

fold reduction in genome replication whereas S225D replicated as wildtype. By confocal 21

microscopy we show that, in the case of S225A, the replication phenotype correlated with an 22

altered subcellular distribution of NS5A. This phenotype was shared by other mutations in the 23

low complexity sequence I (LCS I) – namely S229D, S232A and S235D, but not by mutations 24

with a comparable replication defect that mapped to domain II of NS5A (P315A, L321A). 25

Together with other components of the genome replication complex (NS3, double-stranded 26

RNA and cellular lipids including phosphatidylinositol-4 phosphate), mutant NS5A was restricted 27

to a perinuclear region. This phenotype was not due to cell confluency or another 28

environmental factor, and could be partially transcomplemented by wildtype NS5A. We 29

propose that serine phosphorylation within LCS I may regulate the assembly of an active 30

genome replication complex. 31

IMPORTANCE: The mechanisms by which hepatitis C virus replicates its RNA genome remain 32

poorly characterised. We show here that phosphorylation of the viral non-structural protein 33

NS5A at serine residues is important for the efficient assembly of a complex that is able to 34

replicate the viral genome. This research implicates cellular protein kinases in control of virus 35

replication and highlights the need to further understand the interplay between the virus and the 36

host cell in order to develop potential avenues for future antiviral therapy. 37


.org/D

ownloaded from

http://jvi.asm.org/

3

INTRODUCTION 38

The hepatitis C virus (HCV) currently infects an estimated 170 million individuals worldwide and 39

establishes a chronic infection in 85% of cases, which typically progress to liver cirrhosis and 40

hepatocellular carcinoma (23). The virus has a single stranded RNA genome that codes for a 41

~3000 amino acid polyprotein that is cleaved co- and post-translationally into 10 mature viral 42

proteins; Core, E1 and E2 envelope proteins, which make up the viral particle, followed by the 43

viroporin; p7, and the non-structural 2 (NS2) protein, which possesses autoprotease activity. 44

The remaining NS proteins; NS3, NS4A, NS4B, NS5A and NS5B are necessary and sufficient 45

for genome replication (15). With the advent of the full length clone of a genotype 2a isolate 46

(JFH-1), able to undergo the complete virus lifecycle in cell culture, significant progress has 47

been made in understanding how the different viral proteins contribute to the processes of 48

genome replication and virus assembly. In this regard, NS5A has been shown to play a critical 49

role in both of these processes, as well as being known to perturb numerous host pathways in 50

favour of virus persistence. 51

The 5’ untranslated region (UTR) of the viral genome contains an internal ribosome entry 52

sequence (IRES) that allows viral proteins to be translated in a cap-independent fashion 53

immediately after the viral genome is delivered into the cytoplasm. HCV proteins then rapidly 54

recruit and remodel ER-derived membranes to form a ‘membranous web’ (MW), a sub-cellular 55

structure comprised of single, double and multi-membrane vesicles (SMV, DMV and MMVs) 56

that are enriched in viral (eg NS3, NS5A and NS5B) and host cell (eg VAP-A) proteins (16, 19). 57

By analogy to other positive-sense RNA viruses, this enrichment of viral NS proteins, the 58

presence of double stranded (ds) RNA and the observation that isolated membrane fractions 59

retain replication activity, has led to the MW being proposed as the site of viral genome 60

replication (16). In Huh7 cells, the MW is extensively distributed throughout the cytoplasm, 61

correlating with the observed sub-cellular distribution of the NS5A and NS3 proteins as being 62


.org/D

ownloaded from

http://jvi.asm.org/

4

discrete puncta, throughout the cytoplasm. The formation and maintenance of the MW has 63

been shown to be dependent on the activation of the phosphatidylinositol 4-kinase type III alpha 64

isoform (PI4KIIIα) by NS5A, and the subsequent elevated production of phosphatidylinositol-4-65

phosphate (PI4P) by this lipid kinase (1, 11, 18). Additionally, lipid droplets (LD) are a host 66

organelle responsible for the normal cellular storage of neutral lipids, and have been shown to 67

be central to the HCV lifecycle. While the site of virus assembly has not been elucidated, it is 68

known that both Core and NS5A coat the surface of LD, displacing the host factor adipocyte 69

differentiation-related protein (ADRP). In order for virus genomes that have been newly 70

synthesised in the MW to be packaged into virions, the Core-coated LD are thought to be 71

recruited to the MW, possibly in a NS5A directed manner, in order to allow the encapsidation of 72

viral RNA (4, 14). 73

NS5A is a highly phosphorylated protein, comprised of three domains and tethered to 74

membranes by an N-terminal amphipathic helix. Domain I is highly structured and sufficient for 75

NS5A to dimerise, while domains II and III are intrinsically disordered, but have recently been 76

shown to have elements of transient secondary structure throughout (6, 7, 10). These three 77

domains are linked by low complexity sequences (LCS) (26), a serine rich sequence links the 78

first two domains (LCS I) while a proline rich sequence links the last two domains (LCS II). 79

NS5A exists in two different phosphorylated forms that can be readily resolved by SDS-PAGE. 80

These forms, termed the basal and hyperphosphorylated species, are thought to represent 81

different functional pools of NS5A involved in different stages of the virus lifecycle, but as yet 82

there is little direct evidence to support this hypothesis. Until recently, the sites of NS5A 83

phosphorylation had not been unambiguously mapped, their locations having been inferred 84

from mutagenesis data. Recently however, we and others, have used mass spectrometry to 85

demonstrate that NS5A, purified from cells actively replicating the HCV subgenomic replicon 86

(SGR), is extensively phosphorylated within the LCS I, containing at a minimum seven 87


.org/D

ownloaded from

http://jvi.asm.org/

5

phosphorylation sites (9, 13, 21). We also showed that NS5A contains phosphorylation sites in 88

domain I, domain II and LCS II. Previous work to characterise the role of LCS I phosphorylation 89

has focused almost entirely on a role in virus replication and/or assembly/release, and 90

numerous studies have shown several serines to be important for viral replication, but in a 91

genotype dependent manner. These studies have however, not explored the mechanism 92

through which the phosphorylation of LCS I might be acting. 93

We show here that the phosphorylation of serine 225 within LCS I is critical for the correct 94

localisation of viral proteins during infection. Ablation of serine 225 phosphorylation results in a 95

10-fold reduction in replicative fitness. We show using confocal microscopy that this phenotype 96

is concomitant with a dramatic relocalisation of both cellular and viral factors known to 97

participate in genome replication. This phenotype is shared with other serine mutations in LCS 98

I, further illustrating the complexity of phosphorylation in the region of NS5A. 99

100

MATERIALS AND METHODS 101

Plasmids. DNA constructs of and the full length pJFH-1 virus (27) and the luciferase sub-102

genomic replicon (SGR-luc-JFH-1) (25) were used throughout. Previously, the unique restriction 103

sites BamHI/AfeI flanking NS5A were introduced into both full length virus and SGR-luc-JFH-1 104

constructs, these constructs were denoted as mJFH-1 and mSGR-luc-JFH-1 respectively, and 105

shown to have no effect on virus genome replication or assembly and release (8). Mutagenesis 106

of NS5A was performed on a LITMUS28i (NEB) subclone containing an NsiI/HindIII fragment of 107

the mJFH-1 cDNA, by Quikchange site-directed mutagenesis (Stratagene), then cloned into 108

either mSGR-luc-JFH-1 or mJFH-1 via flanking BamHI/AfeI sites and confirmed by sequencing. 109

For the insertion of the SNAP/CLIP tag into NS5A, the gene was amplified from the pSNAPf 110

and pCLIPf vectors (NEB) using primers that inserted a flexible linker sequence (GSSGSS) 111


.org/D

ownloaded from

http://jvi.asm.org/

6

either side of the insertion (primers sequences available upon request). This amplicon was 112

inserted into the aforementioned LITMUS28i sub-clone by BclI digestion and correct insertion 113

verified by sequencing. The NS5A(SNAP) or NS5A(CLIP) region was then inserted into the 114

mSGR-luc-JFH-1 by flanking RsrII/AfeI sites. For plasmid expression mutated NS5A coding 115

sequences were subcloned into the pCMV(NS3-5B) vector as described previously (20). All 116

mutations were verified by sequencing. 117

Cell culture. Huh7 cells were cultured in Dulbecco’s Modified Eagles Medium (DMEM; Sigma) 118

supplemented with 10% foetal bovine serum (FBS), 100 IU penicillin/ml, 100 μg streptomycin/ml 119

and 1% non-essential amino acids in a humidified incubator at 37°C, 5% CO2. 120

Electroporation of replicon and virus RNA constructs. The preparation of in vitro 121

transcripts and electroporations for both mSGR-luc-JFH-1 and mJFH-1 were conducted as 122

described previously (20). In brief, 2 x 106 Huh7 cells in DEPC-PBS were electroporated with 5 123

μg of in vitro RNA transcripts using a square wave protocol, 260 V and 25 ms. Subsequently, 124

cells were resuspended in complete DMEM media and seeded at 1 x 104 cells/cm2 culture area, 125

into either 12- or 6-well dishes. 126

Virus titration. Supernatants were titrated on Huh7 cells in 96-well format as follows, clarified 127

virus supernatant were serially diluted 2-fold in complete DMEM media before addition of 100 μl 128

to cells seeded 8 hours prior at 8 x 103 cells/well, final volume 200 μl. The counting of infected 129

cells was automated by acquisition and analysis using an Incucyte ZOOM. Virus titre was 130

determined by averaging the positive cell count from 3 or more adjacent counted wells that 131

gave corresponding titres. 132

SDS-PAGE/Western blot. Cells were washed twice in PBS, lysed in 1 x Glasgow lysis buffer 133

(GLB - 1% (v/v) Triton X-100, 120 mM KCl, 30 mM NaCl, 5 mM MgCl2, 10% (v/v) glycerol, 10 134

mM PIPES-NaOH, pH 7.2 with protease and phosphatase inhibitors) and clarified at 2800 x g 135


.org/D

ownloaded from

http://jvi.asm.org/

7

for 5 min at 4oC before determining and normalising protein concentration by BCA assay 136

(Pierce). Following separation by SDS-PAGE, proteins were transferred to a polyvinylidene 137

difluoride (PVDF) membrane and blocked in 50% (v/v) Odyssey blocking (OB) buffer (LI-COR) 138

in PBS. The membrane was incubated with primary antibodies overnight at 4oC, followed by 139

secondary antibodies for 2 h at RT, both prepared in 50% OB buffer. Primary antibodies used 140

were; anti-NS5A (sheep), 1:5,000 (12); anti-NS3 (sheep), 1:2,000 (in-house); and anti-141

glyceraldehyde-3-phosphate dehydrogenase (GAPDH, mouse), 1:20,000 (Sigma). Secondary 142

antibodies were anti-goat (800 nm) or anti-mouse (700 nm), used at 1:5,000 prior to imaging 143

fluorescence using a LI-COR Odyssey Sa infrared imaging system. Quantification of 144

fluorescently labelled Western blots were carried out using Image Studio v3.1 (LI-COR) using a 145

background subtraction method. 146

Immunofluorescence and confocal microscopy. Cells were washed once in PBS before 147

fixation for 20 mins in 4% (w/v) PFA, cells were subsequently permeabilised in 0.2% (v/v) 148

Triton-X-100, PBS before immunostaining with denoted antibody. Primary antibodies used 149

were; anti-NS5A (sheep), 1:1,000; anti-NS3 (sheep), 1:300 (in-house); anti-dsRNA (mouse), 150

1:200 (Scicons); anti-PI4P (mouse), 1:100 (Echelon). Various fluorescently conjugated 151

secondary antibodies were used at 1:1000 (Life Technology). Note, when conducting 152

immunostaining for dsRNA, all reagents were prepared as ‘RNAase free’ by DEPC treatment. 153

Lipid droplets were stained using the BODIPY(558/568)-C12 dye at 1:1000 (Life Technology) 154

added at the same time as the fluorescent secondary antibody. Confocal microscopy images 155

were acquired on a Carl Zeiss LSM 700 inverted microscope, post-acquisition analysis was 156

conducted using either Zeiss Zen 2012 or Fiji v1.49 software (22). 157

In cell labelling of SNAP/CLIP tag. Huh7 cells harbouring SNAP/CLIP replicons were fixed in 158

4% PFA and permeabilised in 0.2% Triton-X-100, PBS with the addition of 0.9 mM CaCl2 and 159

0.5 mM MgCl2 (PBS++). SNAP and CLIP substrates were dissolved to 1mM in DMSO, before 160


.org/D

ownloaded from

http://jvi.asm.org/

8

preparing a 1μM labelling solution in PBS++ containing 10% (v/v) FBS, and addition to cells for 161

2 h at RT. For immunostaining of NS5A, antibody was added to SNAP/CLIP labelling solution, 162

followed by secondary antibody staining as detailed earlier. 163

Quantification of lipid droplets. For quantification of LD spatial arrangement, images were 164

acquired with the same acquisition parameters, but with variable gain to ensure correct 165

exposure. The spatial coordinates of LD were determined using the FindFoci function of the 166

GDSC plugin for Fiji, with the nuclear envelope being manually outlined (utilising the DAPI 167

staining as reference) and coordinates generated by Fiji. The distance from each lipid droplet to 168

the nuclear envelope was then determined using trigonometry. LD spatial distribution data was 169

generated for 12 randomly selected cells for each JFH-1 virus variant and data combined into a 170

box and whisker plot (Fig. 3D). 171

Statistics. Data sets analysed using a Students t-test assuming a two tailed, unequal variance 172

to determine statistical difference from wildtype (WT). n= 3 or greater throughout and * p<0.05, 173

** p<0.01, *** p<0.001. 174

175

RESULTS 176

Phosphorylation of serine 225 is important for genome replication, but not assembly, of 177

infectious virus. Previously we identified serine 225 as being one of a number of serines within 178

the LCS I of NS5A that is phosphorylated during the normal course of infection. To investigate 179

the role of serine 225 phosphorylation in the virus lifecycle we generated either a 180

phosphoablatant alanine substitution, or a phosphomimetic aspartic acid substitution at this 181

residue, in the context of both mSGR-luc-JFH-1 and mJFH-1 constructs, by site directed 182

mutagenesis as described previously (21). Huh7 cells were then electroporated with in vitro 183

transcripts of mSGR-luc-JFH-1 and genome replication followed over 72 h by luciferase activity. 184


.org/D

ownloaded from

http://jvi.asm.org/

9

As seen previously, the phosphoablatant mutation S225A resulted in a 10-fold reduction in 185

genome replication efficiency, while the phosphomimetic mutation S225D had no discernible 186

effect (Fig. 1A). To investigate whether the phosphorylation of serine 225 had additional effects 187

on the release of infectious virus, in vitro transcripts of the mJFH-1 virus with the S225A/D 188

mutations were electroporated into Huh7 cells and released virus titre followed over 144 h by 189

focus forming assay. Figure 1B shows that the S225A mutant virus has around a 10-fold 190

reduction in infectious virus compared to WT, while the S225D mutant virus had titres 191

indistinguishable from wildtype. This reduction in released virus titre for the S225A mutant 192

correlated with the impairment to replication, in line with previous studies showing that genome 193

replication is the rate limiting step in the release of infectious virus (2). These data illustrate that 194

the phosphorylation of serine 225 is important in the replication of viral genomes, and not the 195

release of infectious virus, as previously concluded. 196

Cell lysates from infected cells were collected at 96 and 144 hpe, normalised for protein 197

concentration and analysed by SDS-PAGE/western blot. It was found that there was a reduction 198

in the level of cellular NS5A to a degree comparable with the reduction in virus replication (Fig. 199

1C), quantification of the western blot confirmed this (Fig. 1D). However, when the ratio of 200

hyper- to basal-phosphorylation was determined it was found that the S225A mutant, at both 201

early and late time points, exhibited a reduction in the extent of hyperphosphorylation (Fig. 1D). 202

This correlated with the involvement of the LCS I in hyperphosphorylation previously reported 203

by us and others (9, 13, 21). 204

It was noted when normalising protein concentration for SDS-PAGE that cells infected with the 205

S225A mutant virus were growing at a faster rate than wildtype and S225D mutant virus 206

infected cells (data not shown). To account for this, the released virus titre at 96 and 144 hpe 207

was normalised to cellular protein concentration, used here as a surrogate for cell number, and 208

it was observed that the S225A mutant still exhibited an approximate 10-fold reduction in virus 209


.org/D

ownloaded from

http://jvi.asm.org/

10

release compared to wildtype (Fig. 1E). To ensure that cell confluency (known to negatively 210

affect HCV replication) was not affecting released virus titres, electroporated cells were also 211

passaged at 96 hpe, and at 144 hpe released virus titre determined and normalised to cellular 212

protein (Fig. 1E). As previously, the S225A mutant exhibited a 10-fold reduction in virus release. 213

Ablation of serine 225 phosphorylation alters the sub-cellular distribution of NS5A When 214

titrating virus by focus forming assay the sub-cellular distribution of NS5A(S225A) was 215

observed to be markedly different from that of the wildtype protein. Typically in wildtype JFH-1 216

infected cells NS5A was present throughout the cytoplasm in discrete puncta, however in cells 217

infected with the S225A mutant virus NS5A localised into defined, condensed structures. To 218

investigate this further, Huh7 cells were electroporated with in vitro transcripts of mJFH-1 virus 219

(wildtype, S225A and S225D), fixed at 96 hpe and immunostained for NS5A as described. As 220

previously observed, wildtype NS5A was distributed into small, discrete puncta that were 221

abundant and relatively uniformly distributed throughout the cytoplasm (Fig. 2A). In contrast 222

NS5A(S225A) exhibited bright, condensed clusters that were less abundant and typically 223

localised around the nucleus (Fig. 2B). Importantly the phosphomimetic mutation S225D 224

exhibited a wildtype phenotype (Fig. 2C), strongly suggesting that phosphorylation of serine 225 225

is important for the normal sub-cellular localisation of NS5A. 226

The S225A mutation results in relocalisation of genome replication complexes. We then 227

asked whether the relocalisation of NS5A(S225A) resulted in any changes to the distribution of 228

other viral and cellular markers of replication complexes. Firstly, we looked at the distribution of 229

lipid droplets (LD), which serve as a neutral lipid reservoir for the host and have been shown 230

previously to be an essential host factor involved in HCV replication, with both NS5A and core 231

localising to their surface in infected cells (14). In the case of both wildtype and the two mutant 232

JFH-1 viruses, NS5A colocalised to the periphery of LD (Figs. 3A-C), and the distribution of the 233

LDs correlated with that of NS5A. Specifically, in cells infected with the S225A mutant virus, 234


.org/D

ownloaded from

http://jvi.asm.org/

11

LDs clustered close to the nucleus, whereas in the context of either wildtype or S225D, the LDs 235

were more evenly distributed throughout the cytoplasm. In uninfected cells the distribution of 236

LDs more closely resembled the wildtype or S225D infected cells (Fig 3D). 237

To quantify this LD redistribution, we acquired images of 12 randomly selected cells for the 238

wildtype and each virus mutant, and quantified both the number of LDs, and their distance from 239

the nuclear membrane. With this quantitative approach we saw that in comparison to uninfected 240

cells, wildtype virus infected cells exhibited a broader distribution throughout the cytoplasm, 241

whereas in the context of the S225A mutation there was a significant shift of LD towards the 242

nucleus (Fig 3E). Interestingly however, the S225D mutation had an intermediate phenotype 243

that resembled uninfected cells, where there was a slight shift of LD towards the nucleus, 244

indicating that this mutation was not quite fully able to re-capitulate the effect of phosphorylation 245

at serine 225, at least with respect to LD distribution (Fig. 3E). With this data we were also able 246

to interrogate the total number of LD per cell. There was around a 2-fold increase in the number 247

of LD in cells infected with either wildtype or the S225D mutant viruses, compared to naïve cells 248

(Fig. 3E). This was not the case for the S225A mutant, where the number of LD per cell was 249

broadly comparable to uninfected cells (Fig. 3E), suggesting that NS5A(S225A) is unable to 250

increase the formation of LD upon infection. 251

The viral protease/helicase, NS3, is another key component of genome replication complexes, 252

therefore to further investigate the potential redistribution of these complexes by the S225A 253

mutation we co-stained infected cells with antibodies to both NS5A and NS3. As expected, NS3 254

extensively colocalised with NS5A in all cases, such that in the context of NS5A(S225A) it 255

exhibited a restricted perinuclear distribution, whereas for both wildtype and NS5A(S225D), 256

NS3 was extensively distributed throughout the cytoplasm (Fig. 4A). A similar phenotype was 257

observed for the distribution of double stranded (ds) RNA, a hallmark of viral genome replication 258

that can be detected using a specific antibody, J2 (24). As shown in Fig. 4B, in cells infected 259


.org/D

ownloaded from

http://jvi.asm.org/

12

with mJFH-1 the dsRNA antibody detected small, bright puncta, typically of a uniform size and 260

distributed throughout the cytoplasm. In cells electroporated with a replication-defective mutant 261

(GND) however, there was little to no reactivity towards the dsRNA antibody, suggesting that 262

this staining was specific to cells undergoing virus genome replication (Fig 4B). Furthermore, 263

this dsRNA staining could be observed at very early time points post infection (24 h), further 264

supporting the notion that these puncta are active replication complexes, and not accumulation 265

of structured RNA in stress granules (data not shown). Whilst the distribution of dsRNA-positive 266

punctae was not as extensive as that of lipid droplets (Fig. 3) or NS3 (Fig. 4A), it again 267

correlated to the distribution of NS5A. Thus in cells infected with either wildtype or S225D 268

mutant viruses, dsRNA-positive punctae could be seen throughout the cytoplasm, whereas for 269

the S225A mutant virus they were less numerous and restricted to a perinuclear distribution. 270

The limited distribution of the dsRNA-positive punctae may be a consequence of the lack of 271

sensitivity of the antibody, or may indicate that there are only a limited number of active 272

replication complexes, much less than the total number of structures containing NS3 and NS5A. 273

Lastly, we investigated the distribution of a host cell factor, the phospholipid phosphatidylinositol 274

4-phosphate (PI4P). It has been well established that infection by HCV, and many other RNA 275

viruses, results in the increased production of PI4P lipids, and it is widely accepted that this is 276

one of the mechanisms through which HCV is able to induce the formation of the membranous 277

web (1, 11, 18). Using an antibody specific to PI4P we showed that in cells infected with 278

wildtype or S225D mutant viruses, PI4P levels were elevated in comparison to uninfected cells 279

(data not shown), and that there was a strong colocalisation between NS5A and PI4P, with 280

PI4P typically forming small irregular puncta (Fig. 4C). In contrast, cells infected with the S225A 281

mutant virus showed a dramatic reduction in the abundance of PI4P punctae and their 282

relocalisation to a perinuclear distribution along with NS5A (Fig. 4C). 283


.org/D

ownloaded from

http://jvi.asm.org/

13

From this confocal microscopy and immunofluorescence analysis we conclude that the 284

phosphoablatant mutation S225A in NS5A is sufficient to induce the relocalisation of numerous 285

factors associated with HCV genome replication. It is important to note that we did not observe 286

loss of colocalisation of NS5A with these factors, reflecting the fact that NS5A(S225A) retained 287

the ability to replicate and release infectious virus. 288

The relocalisation of NS5A(S225A) and viral/host factors is not a result of reduced HCV 289

replication. It was important to establish that this phenotype was specifically the result of the 290

loss of phosphorylation at serine 225, and not the result of the lower levels of replication (and so 291

viral proteins) that is observed with the S225A mutation. To achieve this we utilised two 292

previous characterised mutations within domain II of NS5A, P315A and L321A, that had been 293

previously shown to have a similar (approx. 10-fold) impairment to virus replication as S225A 294

(20). 295

As expected, both the NS5A(P315A) (Fig 5A), and NS5A(L321A) (data not shown), mutant 296

viruses exhibited an approximately 10-fold reduction in released virus which was comparable to 297

the S225A mutation, confirming that S225A, P315A and L321A exhibited a similar level of 298

impairment in virus replication. Both mutant viruses exhibited a similar reduction in levels of 299

NS5A expression as seen for S225A (Figs 5B and 5C and data not shown), and exhibited a 300

wildtype ratio of NS5A hyper- to basal-phosphorylation (Fig. 5C and data not shown), in 301

comparison to both S225 mutants, consistent with the previously proposed notion that S225 is a 302

site of phosphorylation in the hyperphosphorylated species of NS5A (21). Therefore, if the 303

phenotype of NS5A relocalisation was the result of reduced replication, it would be expected 304

that both P315A and L321A would show the same phenotype. 305

Huh7 cells infected with NS5A(P315A) (Fig 5D), or NS5A(L321A) (data not shown) mutant 306

viruses were analysed by confocal microscopy for the abundance and distribution of NS5A and 307


.org/D

ownloaded from

http://jvi.asm.org/

14

LD. Reassuringly the distribution of NS5A for both these mutants was indistinguishable from 308

wildtype, with no evidence of the restriction to a perinuclear distribution. Quantitative analysis 309

revealed no significant difference between wildtype and the two mutants with regard to the 310

distance of LD from the nuclear membrane (Fig. 5E). Similar results were seen for the 311

distribution of NS3 and PI4P (Fig 5D), as well as for dsRNA (data not shown). It is noteworthy 312

that, although the distribution of PI4P lipids in NS5A(P315A) virus-infected cells was similar to 313

wildtype, there was an obvious reduction in their quantity, suggesting that this mutant might 314

have an independent effect on NS5A-mediated PI4K activation. This observation 315

notwithstanding, we conclude that the effect of the S225A mutation on the distribution of 316

replication complex components cannot be explained by a reduction in replicative fitness, rather 317

it is specifically the loss of phosphorylation at serine 225 that is responsible for this phenotype. 318

The S225A phenotype does not correlate with a loss of hyperphosphorylation. Although 319

our original observation was made during routine virus titration experiments we had not 320

screened an entire panel of serine mutants within LCS I. Given the complexity of the 321

phosphorylation patterns within this region, and the potential role of S146 in regulating 322

hyperphosphorylation (21), we therefore generated a complete panel of either phosphomimetic 323

or phosphoablatant mutants in all 8 serines in LCS I and S146, and examined their phenotypes 324

both by virus titration and immunofluorescence. Consistent with our previous analysis (21), and 325

that of Masaki et al (13), most mutations had no effect on virus titre, however S225A, S232A 326

and S235D all exhibited a significant reduction (approx. 10-fold) (Fig. 6A). S232A and S235D 327

also shared the restricted perinuclear distribution of NS5A and LD (Fig. 6B). As demonstrated 328

previously (13), S229D did not produce any infectious virus however it exhibited limited RNA 329

replication and protein expression (13, 21), and the distribution of both NS5A and LD was again 330

restricted to a perinuclear region. We conclude that phosphorylation of different serines within 331

LCS I can have opposing effects on the subcellular distribution of NS5A. 332


.org/D

ownloaded from

http://jvi.asm.org/

15

Lastly, we examined the phenotype of a double mutant – S146D, S225D – we had previously 333

shown that this mutant replicated as well as wildtype but almost completely abolished 334

hyperphosphorylation (21). The distribution of NS5A in cells infected with this mutant was 335

comparable to wildtype (Supplementary Fig 1), confirming that the S225A phenotype of 336

restricted perinuclear distribution could not be explained by a defect in hyperphosphorylation. 337

As an aside, none of the mutants (including S225A) exhibited any significant differences in 338

sensitivity to DCV compared with wildtype. The EC50 values ranged from 10-57 pM (in our 339

hands a wildtype JFH-1 replicon had an EC50 of 57 pM), and although treatment with DCV 340

reduced the abundance of NS5A it did not alter the widespread cytoplasmic distribution (data 341

not shown). 342

The S225A phenotype is not the result of cell culture conditions. To enable us to 343

interrogate further the S225A phenotype, we utilised the SNAP and CLIP tag (NEB) technology. 344

This complementary tag system allows for specific conjugation of either benzylguanine (SNAP) 345

or benzylcytosine (CLIP) derivatives, typically fluorescently labelled, to the SNAP or CLIP tag 346

respectively. This approach has been recently described for NS5A (5), and we reasoned that it 347

would enable us to differentially label wildtype NS5A, S225A and S225D mutants within the 348

same population of cells. The advantage over traditional GFP- and RFP-tags is a combination 349

of the flexibility offered with the SNAP/CLIP system and the high sequence similarity between 350

the SNAP and CLIP tags. 351

The SNAP/CLIP tags were initially introduced into the wildtype mSGR-luc-JFH-1 replicon (8) at 352

position 431 in NS5A (domain III) as this site has previously been shown to tolerate large 353

insertions (Fig. 7A). We confirmed that the insertion of either tag had no discernible effect on 354

the replication kinetics of the mSGR-luc-JFH-1 replicon when compared to the untagged 355

wildtype replicon (Fig. 7B). As the SNAP and CLIP-tags share a very high sequence similarity, 356

yet retain different substrate specificity, it was important to establish that the insertion of the tag 357


.org/D

ownloaded from

http://jvi.asm.org/

16

at an internal position did not introduce conformation changes that affected the substrate 358

specificity. As such, cells harbouring the mSGR-luc-JFH-1-NS5A(SNAP) replicon were fixed at 359

72 hpe, immunostained for NS5A and labelled with a CLIP-TMR fluorophore. The converse, 360

where NS5A(CLIP) was labelled with the SNAP-TMR fluorophore, was also carried out. Fig. 7D 361

shows that both the SNAP and CLIP tags retain both catalytic activity and specificity towards 362

the appropriate SNAP/CLIP substrate, with SNAP-TMR only labelling the NS5A(SNAP) and 363

CLIP-TMR only labelling the NS5A(CLIP). The corresponding immunostaining for NS5A also 364

showed a very high degree of correlation, confirming that there was no sub-population of either 365

NS5A(SNAP) or NS5A(CLIP) that was not labelled with the SNAP/CLIP substrates. 366

We proceeded to use the SNAP/CLIP system to establish whether there was a hitherto 367

unknown environmental factor that was inducing the S225A phenotype. To do this we 368

generated mSGR-luc-JFH-1-NS5A(S225A-SNAP) and mSGR-luc-JFH-1-NS5A(S225D-CLIP) 369

replicons. As shown in Fig. 7B, these two mutants retained their replication phenotypes (10-fold 370

impairment for S225A and wildtype for S225D) when expressed as SNAP/CLIP fusions 371

Analysis of cell lysates by western blot showed comparable abundance and hyper/basal 372

phosphorylation ratios of NS5A to that observed previously for the untagged mutants (Fig. 7C). 373

As seen in the context of virus infected cells (Figs. 3 and 4), cells harbouring the NS5A(SNAP-374

S225A) replicon exhibited both a reduced abundance, and restricted perinuclear distribution, of 375

NS5A (Fig. 7E)), in comparison to either wildtype (SNAP tagged) or NS5A(S225D-CLIP) 376

replicon harbouring cells. 377

To confirm that the S225A phenotype was not due to an environmental factor such as 378

differences in cell confluency, we electroporated either mSGR-luc-JFH-1-NS5A(CLIP) or 379

mSGR-luc-JFH-1-NS5A(S225A-SNAP) replicon RNAs into Huh7 cells and co-seeded the 380

electroporated cells into single cultures. By confocal microscopy we observed that within the 381

same population of cells, in the same field of view, those harbouring the wildtype NS5A(CLIP) 382


.org/D

ownloaded from

http://jvi.asm.org/

17

replicon had a normal distribution of NS5A, while those harbouring the NS5A(S225A-SNAP) 383

replicon continued to show the restricted NS5A localisation (Fig. 8A). 384

To further investigate the NS5A(S225A) phenotype we utilised the SNAP/CLIP tag technology 385

to ask if NS5A(S225A) could be transcomplemented by wildtype NS5A. In our experience the 386

simultaneous introduction of two replicons into a single cell is a rare event, for reasons that are 387

not clear, particularly given the ratio of RNA molecules to cells in a typical electroporation 388

experiment. However, after co-electroporation of mSGR-luc-JFH-1-NS5A(CLIP) and mSGR-389

luc-JFH-1-NS5A(S225A-SNAP) replicon RNAs into Huh7 cells we did observe some cells that 390

were stained with both SNAP-505 and CLIP-TMR fluorophores, consistent with the presence of 391

both actively replicating replicons in a single cell. In these cells we observed two distinct 392

phenotypes; either the presence of wildtype NS5A(CLIP) was sufficient to restore the 393

localisation of NS5A(S225A-SNAP) to that of wildtype, resulting in a broad distribution of both 394

and high level of colocalisation (Fig. 8B); or the distribution of NS5A(S225A-SNAP) remained in 395

bright, condensed and amorphous structures, with wildtype NS5A(CLIP) being distributed 396

normally throughout the same cell (Fig. 8C). 397

398

DISCUSSION 399

We present here our investigation of the phenotype of mutations at residue serine 225 in NS5A 400

of JFH-1. We propose that the phosphorylation of serine 225 is important for genome 401

replication; mutation of this residue to alanine (S225A) resulted in a relocalisation of NS5A and 402

other components of the HCV genome replication complex (NS3, dsRNA and PI4P lipids) to a 403

perinuclear distribution, in comparison to either wildtype or a phosphomimetic (aspartic acid) 404

substitution (S225D). This altered sub-cellular localisation was accompanied by a 10-fold 405

reduction in genome replication, again in comparison to wildtype or S225D. This phenotype, 406


.org/D

ownloaded from

http://jvi.asm.org/

18

together with the reduction in genome replication, was shared with three other mutants in LCSI 407

(S229D, S232A and S235D) but importantly was not seen in the context of other NS5A mutants 408

with a similar replication defect (P315A and L321A). These data are consistent with serine 409

phosphorylation in LCSI either positively (S225 and S232), or negatively (S229 and S235), 410

regulating the formation and distribution of replication complexes. This conclusion also fits well 411

with the concept of a phosphorylation-mediated switch of NS5A functions between replication 412

and virus assembly, although the mechanics of such a switch remain elusive. Intriguingly, S232 413

corresponds to the residue which was the target for a dramatic culture adaptation (20,000-fold) 414

in the context of genotype 1b replicons (S2204I in polyprotein numbering). Despite the 415

conservation of the sequence of LCS I between HCV genotypes these observations suggest 416

that this does not manifest in a concomitant conservation of function as both we (Fig 6A), and 417

others (13) do not see any major effect of mutations at this residue on JFH-1 virus replication. In 418

the future it will be important to compare the sites of phosphorylation and phenotypes in a range 419

of HCV genotypes to address these questions. 420

How might the altered distribution of the mutant NS5A be explained? A clue may come from 421

the experiments examining the distribution of PI4P lipids (Fig 4C). PI4P lipids are produced by 422

PI4K, and previous work from other laboratories (1, 11, 18) has demonstrated that the PI4KIIIα 423

isoform is required for HCV genome replication. Indeed PI4KIIIα has been shown to bind to 424

NS5A, the binding site mapping to a short region at the C-terminus of domain I (residues 202-425

210) (17), and disruption of this interaction had a profound effect on replication (100-fold 426

reduction), blocked the stimulation of PI4P formation, and led to a collapse of the membranous 427

web. In this regard, the restricted distribution and clustering of NS5A seen in the context of the 428

S225A mutation is reminiscent of the NS5A localisation in cells silenced for expression of 429

PI4KIIIα (18). We propose therefore that the replication defect of S225A may in part be 430

explained by the inability of this NS5A mutant protein to effectively recruit and activate PI4KIIIα. 431


.org/D

ownloaded from

http://jvi.asm.org/

19

This simplistic explanation is complicated by the results of experiments in which we expressed 432

NS5A in the context of an NS3-5B polyprotein from a CMV promoter driven plasmid vector 433

(data not shown). In this situation the non-structural proteins can form a replication complex but 434

in the absence of a cognate RNA template no genome replication can occur. Intriguingly, under 435

these conditions the S225A phenotype is not apparent: the distribution and abundance of both 436

NS5A and PI4P in transfected cells is very similar for both wildtype and S225A. Thus we would 437

further propose that the function of S225 phosphorylation in regulating interactions between 438

NS5A and PI4KIIIα (and/or other cellular components) is only manifest in the context of a 439

replication complex that is actively replicating the viral genome. What is difficult to reconcile is 440

the observation that inhibiting the interaction between NS5A and PI4KIIIα increased 441

hyperphosphorylation of NS5A (17), concomitant with a reduction in virus genome replication. 442

In contrast, our data show a loss of hyperphosphorylation of S225A (Fig. 1D) and indeed our 443

previous data (21) failed to observe a correlation between hyperphosphorylation and replication. 444

This discrepancy serves to highlight the complexity of NS5A phosphorylation, particularly 445

regarding the multiple phosphorylation events within LCS I. There is a need to fully 446

characterise both the kinases involved and the consequences of phosphorylation to clarify the 447

function of these post-translational modifications in the virus lifecycle. 448

An alternative explanation for the restricted distribution of the S225A mutant NS5A is that it was 449

impaired in its ability to traffic around the cytoplasm. In an elegant set of experiments using 450

both tetracysteine and GFP-tagged NS5A containing viruses, it was shown recently that NS5A 451

was present in two classes of motile structures – the majority were relatively static but some 452

exhibited rapid, long-range and sporadic movement throughout the cytoplasm (5). The rapid 453

movement was shown to involve the microtubule network and was dynein-dependent. It is thus 454

formally possible that NS5A(S225A) is impaired in interactions with microtubules or dynein, 455

preventing its movement around the cytoplasm and leading to its accumulation in the 456


.org/D

ownloaded from

http://jvi.asm.org/

20

perinuclear region. However, we do not believe this is the case because live cell imaging of 457

either wildtype or S225A/D mutant viruses in which NS5A was Emerald-GFP (emGFP)-tagged 458

did not reveal any gross differences in the motility of NS5A positive structures (Supplementary 459

video). In addition, a defect in dynein dependent trafficking might intuitively be expected to 460

result in a distribution around the periphery of the cell, as dynein moves towards the minus end 461

of microtubules which is usually at the cell centre. 462

The use of the SNAP/CLIP tagging system allowed us to differentially label NS5A(S225A-463

SNAP) and wildtype NS5A(CLIP) within one population of cells, thereby visualising in the same 464

field of view, cells that were either harbouring S225A and wildtype replicons. Importantly this 465

analysis demonstrated that the distribution of NS5A(S225A) was still markedly different from 466

wildtype (Fig. 7E), illustrating that the altered localisation of NS5A was not due to an 467

environmental factor. However, the SNAP/CLIP system also revealed that the presence of 468

wildtype NS5A was able to complement the S225A defect in only a proportion of cells . This 469

suggests that additional, as yet undefined, stochastic factors such as stage of cell cycle or 470

availability of cellular factors, may contribute to the observed phenotype. 471

In conclusion, this study highlights an important role of phosphorylation in regulating the 472

function of NS5A. Clearly there is much to do to define the mechanisms underpinning this 473

effect, in particular we need to understand the molecular interactions that are modulated by 474

phosphorylation of NS5A at serine 225 and other serines in LCS I. Whilst a number of 475

candidates such as PI4KIIIα are already apparent, there are many other NS5A interacting 476

proteins that should be considered and a proteomics approach to compare the interactome of 477

both wildtype and mutant forms of NS5A is an appropriate way forward. Such experiments are 478

currently underway in our laboratory. 479

480


.org/D

ownloaded from

http://jvi.asm.org/

21

ACKNOWLEDGEMENTS. 481

We thank John McLauchlan (Centre for Virus Research, University of Glasgow) for the SGR-482

JFH-1-luc construct, Takaji Wakita (National Institute for Infectious Diseases, Tokyo) for pJFH-483

1, Thomas Pietschmann (Hannover) for the NS3 antibody, and Joseph Shaw (Leeds) for help 484

with the daclatasvir EC50 analysis. Finally, we thank Volker Lohmann and Dave Rowlands for 485

insightful comments on this work. This work was funded by a Wellcome Trust Senior 486

Investigator Award to MH (096670MA). JM is a Royal Society University Research Fellow 487

(grant number UF100419), the confocal microscope used for live cell imaging of HCV infected 488

cells was funded by a Royal Society equipment grant (grant number RG110306). 489

490

FIGURE LEGENDS. 491

Figure 1. Characterisation of the phosphorylation site serine 225. (A) Huh7 cells were 492

electroporated with in vitro transcripts of the indicated mSGR-luc-JFH-1 genomes, seeded into 493

96-well format and incubated for 4, 24, 48 and 72 h post-electroporation (hpe). Luciferase 494

activity was determined at 24, 48 and 72 hpe, and normalised to the 4 hpe value. (B) Huh7 cells 495

were electroporated with in vitro transcripts of the indicated mJFH-1 virus genomes, seeded into 496

6-well format and cell supernatant harvested at denoted time point, released virus titre was 497

subsequently determined by focus forming assay. (C) At 96 and 144 hpe cells from (B) were 498

lysed and analysed by SDS-PAGE/western blot. (D) Total NS5A levels and the ratio of hyper- to 499

basal-phosphorylation were quantified from fluorescent western blots. (E) In order to account for 500

the differences in cell growth rate the released virus titre was normalised to total protein, cells 501

were also passaged 1:4 at 96 hpe before harvesting released virus supernatants at 144 hpe. 502

Figure 2. Sub-cellular distribution of wildtype and serine 225 mutant NS5A in JFH-1 503

infected cells. Huh7 cells were electroporated with in vitro transcripts of either wildtype mJFH-504


.org/D

ownloaded from

http://jvi.asm.org/

22

1 (A), NS5A(S225A) (B) or NS5A(S225D) (C) mutants, cells were seeded onto coverslips and 505

incubated for 96 h prior to fixation, immunostaining for NS5A (sheep, 1:1000) and imaging by 506

confocal microscopy. 507

Figure 3. Analysis of lipid droplet redistribution in infected cells. Huh7 cells were 508

electroporated with in vitro transcripts of either wildtype mJFH-1 (A), NS5A(S225A) (B) or 509

NS5A(S225D) (C) mutants, cells were seeded onto coverslips and incubated for 48, 72 and 96 510

hpe, prior to fixation, immunostaining for NS5A and LDs and imaging by confocal microscopy. 511

Uninfected control cells were also processed at 96 h post-seeding. Spatial data for LD was 512

determined from 12 cells for each virus using the GDSC plug-in for Fiji, this data was used to 513

determine the number of LD per cell as well as the distance of each LD from the nuclear 514

envelope (E), box-and-whisker plots show the 10-90% distribution. ***, P< 0.05, significant 515

difference from wildtype. 516

Figure 4. Colocalisation of NS5A phospho-mutants with host and viral factors. Huh7 cells 517

were electroporated with in vitro transcripts of either wildtype mJFH-1, NS5A(S225A) or 518

NS5A(S225D) mutants, and seeded onto coverslips. At 96 hpe cells were fixed and 519

immunostained for NS5A and either NS3 (A), dsRNA (B) or the phospholipid, PI4P (C). An 520

NS5B GND (non-replicating) control is included for the dsRNA staining. Colocalisation scatter 521

plots are of a 200 x 200 μm area containing >10 cells, and show the correlation of absolute 522

values. The correlation value, r, is shown for each image. 523

Figure 5. An NS5A domain II mutation that impairs replication does not have the same 524

redistribution phenotype as NS5A(S225A). Huh7 cells were electroporated with in vitro 525

transcripts of either wildtype mJFH-1, NS5A(S225A), NS5A(S225D) or NS5A(P315A) mutants . 526

(A) Released virus titre was determined at 96 hpe by focus forming assay. Virus titre was 527

normalised to cellular protein in order to account for differences in cell growth rate. (B) At 96 528


.org/D

ownloaded from

http://jvi.asm.org/

23

hpe cells were lysed and analysed by SDS-PAGE/Western blot for NS5A and GAPDH. (C) 529

Total NS5A levels and the ratio of hyper- to basal-phosphorylation were quantified from 530

fluorescent western blots. (D) At 96 hpe, electroporated cells were fixed and immunostained for 531

NS5A and either LD, PI4P or NS3 as described in Figure 4. (E) Spatial data for LD was 532

determined from 12 cells for each virus using the GDSC plugin for Fiji, box-and-whisker plots 533

show the 10-90% distribution. ***, P< 0.05, significant difference from wildtype. 534

Figure 6. Other serine mutations in LCSI share the condensed redistribution phenotype 535

with S225A. (A) Huh7 cells were electroporated with in vitro transcripts of either wildtype 536

mJFH-1 or the indicated mutants . Released virus titre was determined at 96 hpe by focus 537

forming assay. (B) At 96 hpe, electroporated cells were fixed and immunostained for NS5A and 538

LD as described in Figure 3. Only the three mutants that shared the phenotype are shown here 539

with wildtype for comparison. See Supplementary Figure 1 for complete data set. (C) Summary 540

table of mutant phenotypes (D) Amino acid sequence of LCSI. * indicates serine residues with a 541

redistribution phenotype. 542

Figure 7. Generation and validation of mSGR-luc-JFH-1 replicons containing either 543

NS5A-SNAP or NS5A-CLIP fusions. (A) Schematic of the mSGR-luc-JFH-1 replicon showing 544

the region of NS5A where the SNAP/CLIP tag was inserted, along with the flanking, flexible 545

linker sequences. JFH-1 NS5A amino acid numbering. (B) Huh7 cells were electroporated with 546

the indicated mSGR-luc-JFH-1 replicon RNAs, and replication followed over 72 h by 547

determining luciferase activity, shown normalised to the 4 hpe. (C) Cells were lysed at 72 hpe 548

and analysed by SDS-PAGE/Western blot for NS5A and GAPDH. A non-specific band (ns) is 549

seen at equal levels in all lysates (D). Huh7 cells electroporated with the NS5A-SNAP (left 550

panel) or NS5A-CLIP (right panel) replicon RNAs were also seeded onto coverslips, and at 72 551

hpe were fixed and immunostained for NS5A, as well as labelled with either SNAP-TMR or 552

CLIP-TMR before imaging by confocal microscopy. (E) Huh7 cells electroporated with the 553


.org/D

ownloaded from

http://jvi.asm.org/

24

indicated mSGR-luc-JFH-1 replicon RNAs were seeded onto coverslips, fixed at 96 hpe and 554

immunostained for NS5A. 555

Figure 8. The redistribution of NS5A(S225A) is not the result of an environmental factor 556

and can be partially transcomplemented by wildtype NS5A. (A) Huh7 cells electroporated 557

with either mSGR-luc-JFH-1-NS5A(CLIP) or mSGR-luc-JFH-1-NS5A(S225A-SNAP) RNAs 558

were co-seeded onto the same coverslip, and at 96 hpe cells were fixed and labelled with either 559

SNAP-505 (green) or CLIP-TMR (red) before imaging by confocal microscopy. (B, C) In vitro 560

transcripts of mSGR-luc-JFH-1-NS5A(CLIP) and mSGR-luc-JFH-1-NS5A(S225A-SNAP) were 561

pooled prior to coelectroporating into Huh7 cells and seeding onto coverslips. At 96 hpe cells 562

were fixed and labelled with SNAP-505 and CLIP-TMR prior to imaging by confocal microscopy. 563

(B) Example of a population of Huh7 cells showing colocalisation between wildtype NS5A(CLIP) 564

and NS5A(S225A-SNAP). (C) Example of an Huh7 cell showing a difference in cellular 565

localisation between NS5A(CLIP) and NS5A(S225A-SNAP). Colocalisation scatter plots are of 566

individual cells, and show the correlation of absolute values. 567

568

REFERENCES 569

1. Berger, K. L., S. M. Kelly, T. X. Jordan, M. A. Tartell, and G. Randall. 2011. Hepatitis 570 C virus stimulates the phosphatidylinositol 4-kinase III alpha-dependent 571 phosphatidylinositol 4-phosphate production that is essential for its replication. J Virol 572 85:8870-8883. 573

2. Binder, M., N. Sulaimanov, D. Clausznitzer, M. Schulze, C. M. Huber, S. M. Lenz, J. 574 P. Schloder, M. Trippler, R. Bartenschlager, V. Lohmann, and L. Kaderali. 2013. 575 Replication vesicles are load- and choke-points in the hepatitis C virus lifecycle. PLoS 576 Pathog 9:e1003561. 577

3. Boulant, S., M. W. Douglas, L. Moody, A. Budkowska, P. Targett-Adams, and J. 578 McLauchlan. 2008. Hepatitis C virus core protein induces lipid droplet redistribution in a 579 microtubule- and dynein-dependent manner. Traffic 9:1268-1282. 580

4. Counihan, N. A., S. M. Rawlinson, and B. D. Lindenbach. 2011. Trafficking of 581 hepatitis C virus core protein during virus particle assembly. PLoS Pathog 7:e1002302. 582


.org/D

ownloaded from

http://jvi.asm.org/

25

5. Eyre, N. S., G. N. Fiches, A. L. Aloia, K. J. Helbig, E. M. McCartney, C. S. McErlean, 583 K. Li, A. Aggarwal, S. G. Turville, and M. R. Beard. 2014. Dynamic imaging of the 584 hepatitis C virus NS5A protein during a productive infection. J Virol 88:3636-3652. 585

6. Feuerstein, S., Z. Solyom, A. Aladag, A. Favier, M. Schwarten, S. Hoffmann, D. 586 Willbold, and B. Brutscher. 2012. Transient Structure and SH3 Interaction Sites in an 587 Intrinsically Disordered Fragment of the Hepatitis C Virus Protein NS5A. J.Mol.Biol. 588 420:310-323. 589

7. Hanoulle, X., A. Badillo, D. Verdegem, F. Penin, and G. Lippens. 2010. The domain 590 2 of the HCV NS5A protein is intrinsically unstructured. Protein and peptide letters 591 17:1012-1018. 592

8. Hughes, M., S. Griffin, and M. Harris. 2009. Domain III of NS5A contributes to both 593 RNA replication and assembly of hepatitis C virus particles. J.Gen.Virol. 90:1329-1334. 594

9. Lemay, K. L., J. Treadaway, I. Angulo, and T. L. Tellinghuisen. 2013. A hepatitis C 595 virus NS5A phosphorylation site that regulates RNA replication. J.Virol. 87:1255-1260. 596

10. Liang, Y., H. Ye, C. B. Kang, and H. S. Yoon. 2007. Domain 2 of nonstructural protein 597 5A (NS5A) of hepatitis C virus is natively unfolded. Biochemistry 46:11550-11558. 598

11. Lim, Y. S., and S. B. Hwang. 2011. Hepatitis C virus NS5A protein interacts with 599 phosphatidylinositol 4-kinase type IIIalpha and regulates viral propagation. J Biol Chem 600 286:11290-11298. 601

12. Macdonald, A., K. Crowder, A. Street, C. McCormick, K. Saksela, and M. Harris. 602 2003. The hepatitis C virus NS5A protein inhibits activating protein-1 (AP1) function by 603 perturbing Ras-ERK pathway signalling. J.Biol.Chem. 278:17775-17784. 604

13. Masaki, T., S. Matsunaga, H. Takahashi, K. Nakashima, Y. Kimura, M. Ito, M. 605 Matsuda, A. Murayama, T. Kato, H. Hirano, Y. Endo, S. M. Lemon, T. Wakita, T. 606 Sawasaki, and T. Suzuki. 2014. Involvement of Hepatitis C Virus NS5A 607 Hyperphosphorylation Mediated by Casein Kinase I-alpha in Infectious Virus Production. 608 J Virol 88:7541-7555. 609

14. Miyanari, Y., K. Atsuzawa, N. Usuda, K. Watashi, T. Hishiki, M. Zayas, R. 610 Bartenschlager, T. Wakita, M. Hijikata, and K. Shimotohno. 2007. The lipid droplet is 611 an important organelle for hepatitis C virus production. Nat.Cell Biol. 9:1089-1097. 612

15. Moradpour, D., F. Penin, and C. M. Rice. 2007. Replication of hepatitis C virus. 613 Nat.Rev.Microbiol. 5:453-463. 614

16. Paul, D., S. Hoppe, G. Saher, J. Krijnse-Locker, and R. Bartenschlager. 2013. 615 Morphological and biochemical characterization of the membranous hepatitis C virus 616 replication compartment. J Virol 87:10612-10627. 617

17. Reiss, S., C. Harak, I. Romero-Brey, D. Radujkovic, R. Klein, A. Ruggieri, I. 618 Rebhan, R. Bartenschlager, and V. Lohmann. 2013. The lipid kinase 619 phosphatidylinositol-4 kinase III alpha regulates the phosphorylation status of hepatitis C 620 virus NS5A. PLoS.Pathog. 9:e1003359. 621

18. Reiss, S., I. Rebhan, P. Backes, I. Romero-Brey, H. Erfle, P. Matula, L. Kaderali, M. 622 Poenisch, H. Blankenburg, M. S. Hiet, T. Longerich, S. Diehl, F. Ramirez, T. Balla, 623 K. Rohr, A. Kaul, S. Buhler, R. Pepperkok, T. Lengauer, M. Albrecht, R. Eils, P. 624 Schirmacher, V. Lohmann, and R. Bartenschlager. 2011. Recruitment and activation 625 of a lipid kinase by hepatitis C virus NS5A is essential for integrity of the membranous 626 replication compartment. Cell Host & Microbe 9:32-45. 627


.org/D

ownloaded from

http://jvi.asm.org/

26

19. Romero-Brey, I., A. Merz, A. Chiramel, J. Y. Lee, P. Chlanda, U. Haselman, R. 628 Santarella-Mellwig, A. Habermann, S. Hoppe, S. Kallis, P. Walther, C. Antony, J. 629 Krijnse-Locker, and R. Bartenschlager. 2012. Three-dimensional architecture and 630 biogenesis of membrane structures associated with hepatitis C virus replication. 631 PLoS.Pathog. 8:e1003056. 632

20. Ross-Thriepland, D., Y. Amako, and M. Harris. 2013. The C terminus of NS5A 633 domain II is a key determinant of hepatitis C virus genome replication, but is not required 634 for virion assembly and release. J.Gen.Virol. 94:1009-1018. 635

21. Ross-Thriepland, D., and M. Harris. 2014. Insights into the complexity and 636 functionality of hepatitis C virus NS5A phosphorylation. J Virol 88:1421-1432. 637

22. Schindelin, J., I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. 638 Preibisch, C. Rueden, S. Saalfeld, B. Schmid, J. Y. Tinevez, D. J. White, V. 639 Hartenstein, K. Eliceiri, P. Tomancak, and A. Cardona. 2012. Fiji: an open-source 640 platform for biological-image analysis. Nat.Met 9:676-682. 641

23. Shepard, C. W., L. Finelli, and M. J. Alter. 2005. Global epidemiology of hepatitis C 642 virus infection. Lancet Infect.Dis. 5:558-567. 643

24. Targett-Adams, P., S. Boulant, and J. McLauchlan. 2008. Visualization of double-644 stranded RNA in cells supporting hepatitis C virus RNA replication. J.Virol. 82:2182-645 2195. 646

25. Targett-Adams, P., and J. McLauchlan. 2005. Development and characterization of a 647 transient-replication assay for the genotype 2a hepatitis C virus subgenomic replicon. 648 J.Gen.Virol. 86:3075-3080. 649

26. Tellinghuisen, T. L., J. Marcotrigiano, A. E. Gorbalenya, and C. M. Rice. 2004. The 650 NS5A protein of hepatitis C virus is a zinc metalloprotein. J Biol.Chem. 279:48576-651 48587. 652

27. Wakita, T., T. Pietschmann, T. Kato, T. Date, M. Miyamoto, Z. Zhao, K. Murthy, A. 653 Habermann, H. G. Krausslich, M. Mizokami, R. Bartenschlager, and T. J. Liang. 654 2005. Production of infectious hepatitis C virus in tissue culture from a cloned viral 655 genome. Nat.Med. 11:791-796. 656

657


.org/D

ownloaded from

http://jvi.asm.org/


.org/D

ownloaded from

http://jvi.asm.org/

2 the establishment of replication complexes. 4 5 6 7 8 9...

Documents