Modulation of west nile virus capsid protein and viral RNA interaction through phosphorylation 3

| Results
77
4.0 CHARACTERIZATION OF CAPSID (C) PROTEIN AND VIRAL RNA
INTERACTION
4.1 Introduction
Due of the positively charged nature, C proteins of the flaviviruses interacts with
viral nucleic acids. The RNA binding ability of Kunjin virus C protein was confirmed
(Khromykh & Westaway, 1996) but the mechanism of C protein and RNA interaction is
still poorly defined. Hence, this study sought to define C protein binding regions on viral
RNA, viral RNA binding regions on C protein and its interaction in cellular environment.

4.2 Defining the capsid-binding region on the viral RNA.
The capsid-binding region has not been specifically defined and this study sought
to elucidate if such a region exists, if at all. Overlapping viral RNA fragments of
approximately 1 kb in length, spanning the entire viral genome, was synthesized from
amplified DNA fragments of the WNV infectious clone (Li et al., 2005). The synthesized
RNA was then biotinylated. Each of these RNA fragments was used as a probe to detect
for capsid-RNA interaction. Table 4 -1 shows the list of RNA fragments synthesized. The
integrity and size of the amplified DNA (Fig. 4-1A) and in vitro synthesized RNA were
checked with gel electrophoresis (Fig 4-1B).
In the Northwestern blot assay, it was found that that all the RNA fragments,
including Fragment 13 and 14, which represented the negative strand of Fragment 1 and
12, respectively interacted with the purified His-C protein (Fig. 4-2). In addition, there
were no apparent differences in the intensity of each band.

| Results
78
Table 4-1 List of RNA fragments synthesized for C protein pull-down assay and
Northwestern blot
Fragment No.
Region of the WNV genome

1
sense 1-1017
2
sense 960-1974
3
sense 1920-2934
4
sense 2880-3894
5
sense 3830-4839
6
sense 4790-5804
7
sense 5758-6772
8
sense 6728-7742
9
sense 7698-8702
10
sense 8655-9669
11
sense 9613-10617
12
sense 10053-11057
13
anti-sense 1-1017
14
anti-sense 10053-11057

| Results
79











Figure 4-1. Gel electrophoresis of amplified DNA fragments and synthesized viral RNA.
(A) Fragments 1-14 (Table 1) were amplified by PCR using West Nile virus infectious
clone as the template. Each of these templates is tagged with a T7 promoter for in vitro
RNA synthesis. The size and integrity of the PCR product is analysed with DNA gel
electrophoresis. (B) Purified PCR products from (A) were used as templates to synthesis
RNA. The integrity and size of the RNA was analysed with denaturing RNA gel
electrophoresis. The numbers correspond to the fragment number in Table 1. M is the
molecular marker for the DNA and RNA gel electrophoresis.

M 1 2 3 4 5 6 7 8 9 10 11 12 13 14
DNA
1 kb
RNA
M 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1 kb
A
B
| Results
80







Figure 4-2. Northwestern blot with the overlapping WNV RNA fragments. Equal
amounts of His-C protein was loaded into each lane and subjected to SDS-PAGE and
transferred onto a nitrocellulose membrane. The membrane was subsequently probed
with each RNA fragment individually (Upper panel). The numbers above corresponded
to the RNA fragment described in Table 4-1. All the RNA fragments including the anti-
sense Fragments (13, 14) are shown to interact with His-C protein with no observable
differences in band intensities. To ensure equal loading, the same volume of His-C
protein was loaded into each well and subjecting it to Western blot analysis with anti-His
antibody (Lower panel).

RNA
His-C
Fragment 1 2 3 4 5 6 7 8 9 10 11 12 13 14
| Results
81
Since it was not obvious which fragment had a greater affinity for the C protein,
each interaction was quantified by allowing the His-C protein and all the RNA fragments
to form complexes in solution and pulling down the RNA with anti-histidine antibody. If
C protein had preference for a certain region of the WNV RNA genome, this region
would be over-represented when the protein was pulled down. The RNA fragments that
were pulled down were then quantified with RT-PCR. It was found that the over-
represented fragments were Fragment 1, 2, 3, 11, 12, 13, 14 (Fig. 4-3) although all
fragments were detected by RT-PCR. It was surprising to find that the C protein had also
pulled down a significant amount of the negative-stranded WNV RNA of Fragment 1 and
12. The difference in the amount of positive- and negative-stranded RNA being pulled
down [region 1-1017 (Fragment 1 and 13) on the WNV RNA genome] was not
significant. However it was clear that Fragment 2, representing the region 960-1974 on
the WNV RNA genome was over-represented by more than two times relative to other
fragments.

| Results
82
RNA Fragment Number
Amount of RNA pulled down (ng)

Figure 4-3. Capsid protein pull-down of WNV RNA fragments. Overlapping RNA
fragments spanning the entire WNV genome including the anti-sense RNA were
synthesized (Table 4-1) and mixed with purified His-C protein or histidine-tagged
domain III (negative control) of the Dengue E protein (EDIII). Subsequently, anti-His
antibodies-conjugated to sepharose beads were used to pull down the proteins. As a
control, RNA fragments alone were mixed with the antibody-conjugated sepharose beads
to ensure that the RNA was not pulled down unspecifically. The beads were then treated
with proteinase K to remove the proteins in the solution and the RNA was extracted from
the mixture with phenol-chloroform. The extracted portion was then processed for
detection with RT-PCR for each RNA fragment with specific primers. Fragments 1, 2, 3,
11, 12, 13 and 14 are overrepresented in the pull down although RT-PCR was able to
detect trace amounts of other RNA fragments. Amongst all the fragments, Fragment 2
was pulled down the most often. Fragment 13 and 14 are anti-sense fragments of
Fragments 1 and 12. This suggests that C protein binding sites on the WNV RNA are
found in the regions of 1 - 2934 and 9613 - 11057 of the WNV RNA genome.



| Results
83
4.3 Defining the West Nile virus (WNV) RNA-binding region on the capsid (C)
protein
Although the Kunjin virus RNA-binding region on the C protein had previously
been identified to be at the amino- and carboxyl-terminal (Khromykh & Westaway,
1996) greater definition of this region can be achieved by using overlapping peptides of
approximately 23-26 amino acids in length. The list of each peptide and charge at pH 7.0
is detailed in Table 4-2.

Table 4-2. List of C protein peptides synthesized and their charge at pH 7.0.
Fragment

No.
Amino acid
position
Charge at pH 7.0
Amino acid sequence
1
1-23
6.9
mskkpggpgknravnmlkrgmpr
2
18-45
5.9
krgmprglsliglkramlslidgkgpir
3
41-64
3.9
kgpirfvlallaffrftaiaptra
4
61-84
4.1
aptravldrwrgvnkqtamkhllsf
5
79-105
8.1
khllsfkkelgtltsainrrstkqkkr
6
101-123
3.9
kqkkrggtagftillgliacaga

In the peptide dot blot assay, each peptide is blotted onto nitrocellulose paper and probed
with 3’ UTR of the viral genome since this was used in the previously reported
experiment (Khromykh & Westaway, 1996). The peptide dot blot assay showed that
peptides that interacted with the biotinylated-WNV RNA strongly were found to be on
the amino- and carboxyl- terminal of the C protein (Fig. 4-4). The peptides relative
| Results
84
intensity of each blot corresponded to the charge of each peptide. As expected the
stronger the positive charge, the greater the intensity of the blot. As a control the His-C
protein was also blotted on to the membrane and it was functionally capable of binding to
viral RNA (Fig 4-4). This assay showed that the strongest RNA binding regions of the C
protein were between amino acids 1-23 and 79-105. Thus this gives a better definition of
where the RNA binding sites are located (Khromykh & Westaway, 1996).





Figure 4-4. Dot blotting peptides of the C protein to detect interaction with viral RNA.
The peptides listed 1-6 in Table 4-2 as well as purified His-C protein were dot blotted
onto a nitrocellulose membrane. The membrane was dried and blocked with yeast tRNA.
Subsequently, the membrane was incubated with 3’ UTR of the WNV RNA labelled with
biotin conjugated to alkaline phosphatase. The blot was washed and developed with an
alkaline phosphatase substrate. The amount of RNA bound to each peptide corresponds
to the charge of each peptide at pH 7.0 (Table 4-2) and it shows that RNA binding
regions are found on the amino- and carboxyl-terminal.

1 2 3 4 5 6 His-C
| Results
85
4.4 Capsid (C)-RNA interaction in vivo
It has been reported that WNV C protein is imported into the nuclei when a cell is
infected (Bhuvanakantham et al., 2009; Wang et al., 2002). Given that C protein is highly
positively charged and that C protein-RNA interaction has been shown to occur in vitro
(Fig. 4-2), investigation was initiated on how such an interaction would occur in a
cellular environment and if the C protein could translocate viral RNA into the nucleus.
Plasmid encoding myc-tagged C (myc-C) protein was co-transfected with rhodamine-
labelled viral RNA and visualized with either fluorescence or confocal microscopy. It
was shown that, though myc-C protein was localized in the nuclei, the RNA remained in
the cytoplasm (Fig. 4-5A). Confocal microscopy further showed no evidence of viral
RNA being translocated into nucleus by myc-C protein (Fig. 4-5B). Co-localisation of
myc-C protein and synthesized WNV RNA in the nucleus or cytoplasm was not detected
(Fig. 4-5A). This was unanticipated since the highly positively charged nature of the C
protein was expected to interact with the negatively charged RNA.
To eliminate the possibility that the absence of interaction observed in the cells
was not due some inhibitory factors in the cellular environment, it was decided that a
Northwestern blot assay with myc-C protein be performed. The RNA was not able to
bind to myc-C protein immobilized on the nitrocellulose membrane (Fig 4-6). This
demonstrated that the absence of interaction seen in the cellular environment was not due
to due to inhibitory factors in the cells, rather the nature of the myc-C protein. It is
possible that myc-C protein, being expressed in BHK cells, could be phosphorylated. In
contrast, His-C protein was translated in bacteria cells and proteins expressed by
prokaryotic cells are not known to be phosphorylated.

| Results
86
Figure 4-5. Cellular localisation of myc-C protein and synthesized viral RNA in BHK
cells. (A) Visualization of myc-C protein and transfected viral RNA in BHK cells with
fluorescence microscopy. Plasmid encoding myc-C protein and synthesized full length
viral RNA were either transfected individually (i-iv and v-viii) or co-transfected (ix-xii)
into BHK cells and processed for immunofluorescence microscopy 24 hr post
transfection. No co-localisation of myc-C protein (green) and RNA (red) are observed.
Most of myc-C protein is localised in the nucleus (blue) while the RNA is observed in the
cytoplasm. (B) Visualization of myc-C protein and transfected viral RNA in BHK cells
with confocal microscopy. In order to be certain that co-localisation of myc-C protein and
RNA did not occur and that the RNA was not translocated into the nucleus, BHK cells
co-transfected with plasmid encoding myc-C protein and full length viral RNA were
viewed using confocal microscopy. Individual channels showing the nucleus (i), myc-C
protein (ii), RNA (iii) as well as the combined image (iv) are shown. In addition, volume
rendering of the combined image is shown (v). Confocal microscopy shows that the RNA
is not translocated into the nucleus nor is there any observable co-localisation of myc-C
protein and viral RNA. Scale bars are shown on the bottom right of the images.

| Results
87

A
DAPI
RNA
Anti-myc
antibody
Merged
myc-C protein
only
RNA only
myc-C protein
and RNA
i
ii
iii
iv
v
vi
vii
viii
ix
x
xi
xii
| Results
88

B
DAPI
myc-C
RNA
Merged
Volume rendering
i
ii
iii
iv
v
| Results

89














Figure 4-6. Northwestern blot analysis of myc-C protein. Plasmid encoding myc-C
protein was transfected into BHK cells and harvested 24 hour post-transfection. Immuno-
purified (IP) myc-C protein (Lane 1), transfected BHK cell lysate (Lane 2) and
untransfected BHK cell lysate (Lane 3) were subjected to SDS-PAGE and the proteins
were transferred onto a nitrocellulose membrane. The membrane was probed with 3’UTR
RNA of the WNV genome and developed. The membrane was then stripped and probed
with anti-myc antibody. Molecular weight of myc-C protein is slightly less than 17 kDa
and the molecular weight markers are shown on the right. Although the RNA probe is
able to interact with other cellular proteins in the cell lysate (Upper panel, Lanes 2 and 3),
none of these proteins are less than 17 kDa in size, hence the RNA is unable to interact
with myc-C proteins on the membrane.

1 2 3
Anti-myc
antibody
3’ UTR
RNA probe
IP myc-C

Transfected cell lysate
BHK cell lysate
17 kDa
10 kDa
26 kDa
55 kDa
72 kDa
17 kDa
10 kDa
26 kDa
55 kDa
72 kDa
| Results
90
Recognising that the conditions of the previous study were artificially created
since RNA was transfected together with a myc-C protein expressing plasmid, another
assay was developed to study C protein and WNV RNA interaction in vivo. In this assay,
nascent viral RNA was fluorescently labelled in infected cells instead and at the same
viral C protein was probed with anti-C antibodies during the course of an infection. In
order to label nascent viral RNA, Click-IT RNA labelling kit was used. The kit contained
an RNA analog, which was added to the growth media in the cell culture. Hence, the
RNA analog will be incorporated into newly synthesized RNA. However, this labeling
method also labels cellular mRNAs that are transcribed in the nucleus. In order to reduce

the labelling of cellular mRNA, actinomycin D was added to the cells to arrest cellular
transcription in the nucleus. However, high concentrations of actinomycin D is cytotoxic
to the cells, hence an optimal concentration was needed to reduce host mRNA synthesis
without affecting cellular viability. This was achieved by adding a range of actinomycin
D concentrations to BHK cells while observing cellular morphology and RNA labelling.
The optimal concentration of actinomycin D, which reduced cellular RNA labelling
without visible effects of cytotoxicity is 1.0 µM (Fig. 4-7). It can be seen that
fluorescence was detected only in the nucleus and this was consistent with cellular
transcription in the nucleus and that the addition of actinomycin D reduced the number of
cells exhibiting fluorescence.




| Results
91
Figure 4-7. Optimal concentration of actinomycin D needed to disrupt cellular mRNA
synthesis without affecting BHK cell morphology. In order to find the optimal
concentration of actinomycin D needed to disrupt cellular mRNA synthesis in uninfected
cells, a range of actinomycin D concentrations (0.5 - 2.5 µM) were added into the growth
media and nascent RNA was labelled and stained (green) with Click-iT RNA labeling kit
(green). The nucleus of the cell was stained with DAPI. The slides were viewed with
fluorescence microscopy and the fluorescent image was overlaid onto the DIC image of
the slide. RNA labelling of nascent mRNA decreases with increasing concentration of
actinomycin D. At 1.0 µM of actinomycin D BHK cells begin to appear elongated.

| Results
92

0.25 µM actinomycin D
0.5 µM actinomycin D
1.5 µM actinomycin D
1.0 µM actinomycin D
2.5 µM actinomycin D
2.0 µM actinomycin D
| Results
93
With the optimal actinomycin D concentration determined, the efficiency of
labeling viral RNA in infected cells was tested. Cells were infected with WNV, fixed and
processed for fluorescent microscopy at 6 hr, 12 hr, 18 hr and 24 hr post-infection (p.i.).
As a control, mock-infected cells were also labelled for nascent RNA and fixed at 24 hr
p.i. Distinct and punctate green fluorescent dots in the cytoplasm were observed in all
infected cells but the number and intensity of the fluorescent dots increased over time
[Fig. 4-8 (white arrows)]. At 18 hr and 24 hr post infection, diffused cytoplasmic staining
was observed [Fig. 4-8 (red arrows)], whereas no distinct punctate dots in cytoplasm
were observed in the mock-infected cells. Hence, it was concluded that the distinct
punctuate green fluorescent dots found in cytoplasm of the cells were sites of viral RNA
synthesis and at later time points (18 hr and 24 hr) viral RNA had begun to diffuse into
the cytoplasm. Because the RNA labelling kit also labels host RNA, fluorescently

labelled RNA in the nuclei were excluded as theses were sites of host mRNA synthesis.
To observe C protein and viral RNA localisation in infected cells, the cells were
infected with WNV and fixed at 6 hr, 12 hr, 18 hr and 24 hr p.i. Capsid protein in
infected cells was detected with anti-C protein antibodies. Anti-C antibodies were
subsequently stained with anti-mouse antibody conjugated to Alexafluor 594. It was
observed that while the C protein was stained mostly in the nuclei at 6 hr p.i. [Fig 4-9A
(ii)], it was found in the cytoplasm and nucleus at later time points [Fig. 4-9A (iii-v)].
Although in some cells at 12 hr post-transfection, the RNA and C protein were localized
in the cytoplasm, co-localisation of cytoplasmic RNA and C protein was only apparent
in cells at 18 hr and 24 hr post-infection [Fig. 4-9A (iv-v)]. A 3-dimensional rendering of
Fig. 4-9A (iv-v) showed that there were distinct co-localisation of the C protein and viral
| Results
94
RNA in the cytoplasm [Fig. 4-9B (white arrows)]. Hence, in this assay some co-
localisations between the C protein and nascent viral RNA were observed only at the later
time points of an infection.

| Results
95
Figure 4-8. Labelling of viral RNA in infected cells. Cells were infected with WNV and
RNA was labelled with Click-iT RNA labeling kit. The cells were fixed and processed
for microscopy at 6 hr (ii), 12 hr (iii), 18 hr (iv) and 24 hr (v) post infection (p.i.). Mock-
infected cells were fixed and processed after 24 hr. Distinct punctuate green fluorescent
dots (white arrowheads) are seen in the cytoplasm of infected cells (ii-v). In addition
diffused cytoplasmic stainings (red arrows) are observed in infected cells at 18 hr (iv) and
24 hr (v) p.i. No distinct green fluorescent dots or diffused cytoplasmic staining is
observed with mock-infected cells (i). Scale bar is shown on the bottom right of the
images.

| Results
96

(i) Mock infected
(ii) 6 hr p.i.
(iii) 12 hr p.i.
(iv) 18 hr p.i.
(v) 24 hr p.i.
(i) Mock Infected
| Results
97
Figure 4-9. Localisation of C protein and viral RNA in infected cells over 24 hr. (A)
BHK cells were either mock-infected (i) or infected with WNV (ii-v) with an M.O.I of 1
and fixed at 6 hr (ii), 12 hr (iii), 18 hr (iv) and 24 hr (v) p.i. Host RNA synthesis were
partially inhibited with actinomycin D and viral RNA was labelled with Click-iT RNA
labelling kit (green). Viral C protein was probed with anti-C antibody and the anti-C
antibody was detected with anti-mouse antibody conjugated to Alexafluor 594. At early
timings (6-12 hr p.i.) the C proteins are predominantly found in the nuclei (ii-iii) while at
later timings (18-24 hr p.i.) C proteins are found predominantly in the cytoplasm. The
arrows indicate co-localisation of viral RNA with C protein. (B) Panels (i) and (iii) show
the 3-dimenisonal rendering for panels (iv) and (v) in (A), respectively. Panels (ii) and
(iv) are blow ups of an area in (i) and (iii), respectively. The white arrows indicate the co-
localisation (yellow) of C protein (red) and and viral RNA (green) in the cytoplasm.

| Results
98
A

10 µm
Nucleus
RNA
Capsid
Merged
6 hr p.i.
12 hr p.i.
18 hr p.i.
24 hr p.i.
Mock-
infected
ii
iii
iv
v
i
| Results
99

i
i
i
B
18 hr p.i.
24 hr p.i.
i
ii
iii
iv
i
ii

iii
iv
| Results
100
4.5 Phospho-peptides and RNA interaction.
Although some co-localisations of C proteins with viral RNAs at 18 hr and 24 hr
p.i. were observed (Fig. 4-9), the difference between His-C and myc-C proteins’ ability to
interact with viral RNA was still unresolved. Not only were both proteins tagged
differently, they were also expressed in different types of cells. The tags on both these
recombinant proteins were small and therefore unlikely to affect any significant
biochemical alterations to the protein. However, myc-C protein expressed in eukaryotic
cells like BHK cells could undergo post-translational modifications. Conversely, His-C

protein, expressed in prokaryotic cells would not have undergone post-translational
modification. A possible post-translational modification on the myc-C protein that might
have affected RNA binding would be phosphorylation. Since phosphate groups are
negatively charged they could neutralize or attenuate the intrinsic positive charge of C
protein. Hence the difference in RNA binding property between His-C (Fig 4-2) and
myc-C (Fig. 4-6) proteins could probably due to phosphorylation.
Bioinformatics analysis revealed 5 putative phosphorylation sites on the WNV C
protein (Sarafend), 4 of which are serine residues (Table. 4-3). Bioinformatics analysis of
the C protein of other flaviviruses including the Wengler and New York strains of WNV
also revealed multiple putative phosphorylation sites (Table 4-3).





| Results
101
Table 4-3. Putative phosphorylation sites on flavivirus C protein





Virus
Accession no.
(GenBank)
No. of Putative
phosphorylation sites
Amino
acid/Sequence

position
West Nile
(Sarafend)
AY688948
5
S26, S36, S83, S99,
T100,
West Nile
(Kunjin/New
York)
D00246/
FJ151394
5
S26, S36, S83, S99-
100
Japanese
Encephalitis
AAQ73507
3
T2, S36, S83
Yellow Fever
ABB69689
4
S22, S24, S97, S102
Dengue 2
AAC40839
6
S24, S34, T58, T71,
S101
Tick-borne

Encephalitis
AAA86870
2
S19, S97