Modulation of west nile virus capsid protein and viral RNA interaction through phosphorylation 5

Besides these two processes, the cellular localisation of C protein needs to be considered since it has been noted by many investigators Bulich & Aaskov, 1992; Mori et al., 2005; Oh et a

7.0 DISCUSSION AND CONCLUSION

Although the entry and egress of WNV have been studied extensively (Chu et al., 2003; Chu & Ng, 2002a; 2004; Chu et al., 2005; Ng et al., 2001), the processes of

assembly are still not well understood Assembly of the virus can be broken down into two stages – encapsidation of the viral RNA by the C protein to form nucleocapsid and envelopment of nucleocapsid by viral envelope protein (Fig 7-1) This study is focused

on the processes that lead to the formation of the nucleocapsid The most important process is the interaction between C protein and viral RNA Oligomerization of C proteins would no doubt also be an important process Besides these two processes, the cellular localisation of C protein needs to be considered since it has been noted by many

investigators (Bulich & Aaskov, 1992; Mori et al., 2005; Oh et al., 2006; Sangiambut et al., 2008) that C protein localises in the nucleus of infected cell but assembly of the virus

occurs in the cytoplasm Hence, it is postulated that the functions of C protein might be modulated to achieve efficient assembly of the nucleocapsid This study accumulated evidence that phosphorylation of WNV C protein acts as such a modulator of C protein function

Figure 7-1 General diagram of flavivirus assembly process This shows the overview of

the replication process of a virus life cycle In the inset, it shows the assembly process of the nucleocapsid, which involves interaction of the capsid protein (hexagon) with RNA (wavy line) and the oligomerization of the capsid protein Once the nucleocaspid is assembled, it translocates into the lumen of the endoplasmic reticulum (ER), where the envelope protein (circle) would wrap itself around the nucleocapsid The upper panel was

reproduced from a review by Mukhopadhyay et al., 2005

In order to investigate the role of phosphorylation in the processes of nucleocapsid assembly, a series of experiments were performed In addition double and triple mutations were introduced at the amino- and carboxyl- terminal of C protein, respectively This is because it was previously found that flavivirus C protein was still functional when either terminal was deleted suggesting redundancy and functional

flexibility (Kofler et al., 2002; Patkar et al., 2007; Schlick et al., 2009) Moreover,

analysis of other flavivirus C protein sequences also revealed that putative phosphorylation sites were found on the amino- and/or carboxyl- terminal (Table 4-3) This suggested that even though the exact residues or sites may not be conserved, the regions of phosphorylation on C protein were consistent In addition, it was also shown that protein kinase C was responsible for phosphorylating C protein (Fig 5-5) This is the first time C protein of a member of the mosquito-borne flaviviruses has been shown to function as a phospho-protein

Although homology modeling was performed to ascertain that mutations introduced did not perturb the conformation of the C protein it could not be extended to include the mutations at positions 99-100 (Fig 5-4) since the crystal structure of WNV C

protein was solved for amino acids 24 to 96 (Dokland et al., 2004) Therefore only 3 of

the 5 mutation sites were shown (Fig 5-4)

As noted earlier, nuclear localisation of flavivirus C protein have been

demonstrated by various investigators (Bulich & Aaskov, 1992; Mori et al., 2005; Oh et al., 2006; Sangiambut et al., 2008; Wang et al., 2002) and this seemed to be important for the replication of the virus during the earlmmmmy phase [(Bhuvanakantham et al.,

2009), (Fig 7-2)] It was recently discovered that importin-α protein facilitated the

nuclear localisation of Dengue and WNV C proteins (Bhuvanakantham et al., 2009) The

strength of interaction between importin-α protein and WNV C protein was demonstrated

to be enhanced by C protein phosphorylation (Bhuvanakantham et al 2010)

The observation that the cellular distribution of mutant myc-C proteins was more diffused than wild type myc-C protein (Fig 5-7A) is consistent with previous finding observed with Hepatitis C virus (Lu & Ou, 2002) where phosphorylation enhanced nuclear localisation of C protein To complement studies using hypophosphorylated mutant C proteins, protein kinase C activators and inhibitors were used to enhance or attenuate wild type C protein phosphorylation (Fig 5-5) As expected, nuclear localisation of wild type myc-C protein was disrupted in bisindolylmaleimide-treated (PKC inhibitor) cells (Fig 5-8A) The proportion of cells with wild type myc-C protein localized in the nucleus and cytoplasm in bisindolylmaleimide-treated cells was about the same as cells transfected with mutant S26/36/36/83/99/T100A (Fig 5-7B) Because the results from both studies (mutagenesis and PKC inhibitor) were consistent, the phenotype observed in the mutants was due to hypophosphorylation of myc-C protein

Nuclear localisation of C protein is an integral part of virus replication and the inhibition of nuclear localisation of C protein by deleting or mutating the nuclear localisation signal on the WNV or Dengue virus C protein had deleterious effects on the

viruses (Bhuvanakantham et al., 2009) There are suggestions that nuclear localisation of viral C proteins may be a strategy used by viruses to disrupt cytokinesis (Hiscox et al.,

2001; Ning & Shih, 2004) In the case of flaviviruses, the C protein was shown to be

involved in cell-cycle arrest (Helt & Harris, 2005; Oh et al., 2006) and there was some

evidence to suggest correlation between WNV C protein nuclear localisation and cell

cycle arrest (Oh et al., 2006) Nonetheless, whatever the function of flavivirus C protein

nuclear localisation may be, disruption or inhibition of this process was detrimental to

flavivirus replication (Bhuvanakantham et al., 2009; Mori et al., 2005; Sangiambut et al.,

2008) Hence the reduced viral yield of the mutant viruses (Fig 6-3) could be partially explained by the cellular localisation of C protein during early infection

In addition to the above-suggested functions of C protein in the nucleus, nuclear localisation of C protein during the early phase of infection (Fig 6-12, v and vi) may also have another function Translocation of C protein into the nucleus in the first 12 hr

of infection may prevent premature C protein and RNA interaction since viral RNA synthesis occurs in the cytoplasm (Fig 4-8) Investigations into how C protein and viral RNA interacted in a cellular environment showed that there was no co-localisation between myc-C and viral RNA since the RNA remained in the cytoplasm while the myc-

C protein was translocated in the nucleus (Fig 4-5) In addition, wild type C protein and viral RNA were shown to localise in the nucleus and cytoplasm of infected cells, respectively (Fig 4-9) It was only at 18 and 24 hr post-infection that any co-localisation

of C protein and viral RNA was observed in the cytoplasm (Fig 4-9) This showed that C protein did not remain in the nuclei permanently but the location between the nuclei and cytoplasm could be dynamic and regulated These observations coincided with the gradual dephosphorylation of C protein in infected cells (Fig 5-10)

Having observed the presence of mutant C protein in the cytoplasm of infected cells as early as 12 hr post-infection (Fig 6-12, i and ii) and that His-C protein could interact with negative strand viral RNA (Fig 4-2 and Fig 4-3) there is a real possibility

that mutant C protein could have incorporated negative-sense viral RNA into the virions during early replication phase

To investigate if mutant virus would indeed incorporate relatively more sense viral RNA into the virions, viruses were purified for analysis with RT-PCR The results indicated that on average, there was about 10 times more positive strand RNA than negative strand RNA in a pool of wild type virus (Fig 6-14 A) In contrast, the ratio between the positive- and negative-strand RNA in the pool of mutant virus is almost 2:1 (Fig 6-14A) This means that almost 1/3 of the viruses in the mutant pool were non-infectious and this could explain the reduced titre seen in the mutant virus infections (Fig 6-8) The detection of negative strand RNA in a pool of positive strand virus is not all

negative-that surprising On average, for viruses in the Flaviviridae family, the ratio between positive- to negative-strand viral RNA in infected cells is about 10:1 (Quinkert et al., 2005; Richardson et al., 2006) and similar results was also found infected cell lysates

from this study (Fig 6-14A), hence it would be logical to assume that the ratio of positive

to negative strand viral RNA packaged into wild type virions would also be about 10:1

Similar changes in the ratio of genomic and anti-genomic RNA being incorporated into virions due to mutations or deletions of viral C protein have been

reported in the Sendai virus, a single-stranded negative-sense RNA virus (Irie et al.,

2008) The change in the ratio between genomic and anti-genomic RNA packaged into the virions was due to aberrant changes in RNA synthesis However, in the case of WNV, the mutations introduced into the C protein did not cause the ratio of positive- to negative-strand RNA to change in infected cells (Fig 6-14A) Therefore the changes in

the ratio of positive to negative strand RNA being packaged into the virions is likely due

to the relative abundance of the negative-sense RNA during the early phase of infection

Hence the nuclear localisation of C protein could indirectly prevent premature C protein and RNA association, although direct regulation of C protein and RNA interaction was achieved through phosphorylation of C protein Phosphorylation of C protein has been demonstrated in both RNA and DNA viruses to aid the release of

nucleic acid into infected cells and regulate packaging of its genome (Gazina et al., 2000; Ivanov et al., 2001; Law et al., 2003)

To investigate C protein and RNA interaction and how phosphorylation might

affect this interaction in vitro, interaction between C protein and viral RNA needs to be

characterized Although the delineation of RNA binding regions on the WNV C protein is not new (Khromykh & Westaway, 1996), the use of overlapping peptides to map out the RNA binding regions is novel since it gave a better definition as to where the RNA binding sites on the C protein are (Fig 4-4) It was found that the strongest RNA binding regions corresponded with the positive charge of each peptide (Table 4-2) From previous

studies (Khromykh & Westaway, 1996; Patkar et al., 2007) it was shown that 3’ and 5’

UTR of the viral RNA were able to bind to WNV C protein but those were not the only regions that could interact with C protein This study showed that all regions of the viral RNA could interact with His-C protein (Fig 4-2)

In some cases, C protein of other flaviviruses could even interact with

single-stranded DNA to form nucleocapsid particles in vitro (Kiermayr et al., 2004) Hence

interaction of C protein with nucleic acid may be indiscriminate Nonetheless, the strongest binding RNA fragment was confirmed when His-C protein pull-down of viral

RNA was performed in conjunction with RT-PCR (Fig 4-3) From the results in Fig 4-3,

it showed that the encapsidation signal was found in Fragment 2, which is from position 960-1974

What determines C and viral RNA interaction is unclear but it was speculated that secondary RNA structures on the viral could mediate C protein and RNA interaction (Khromykh & Westaway, 1996) The finding that C protein was able to pull down

negative-sense viral RNA in vitro (Fig 4-3, Fragment 13 and 14) reinforced the idea that

C protein and viral RNA interaction may be indiscriminate Hence suggesting that a mechanism such as phosphorylation could be employed by the virus as a form of regulation to ensure a more specific interaction between C protein and viral RNA

Indeed, preliminary investigations with phosphorylated peptides of the RNA binding regions of C protein show that phosphorylation attenuated RNA binding (Fig 4-10) In light of this, the putative phosphorylation sites located within the RNA binding regions of the C protein (Table 4-3 and Fig 4-4) did not seem to be coincidental

To study the effect of phosphorylation of C protein on RNA interaction, Northwestern blot was employed Although this method is typically used to detect protein-RNA interaction, the RNA pull-down assay was also developed in this study to complement the Northwestern blot In the Northwestern blot assay, myc-C protein did not show any interaction with viral RNA (Fig 5-6A) In contrast, using the RNA pull-down assay, the wild type myc-C showed evidence of interacting with RNA since it was pulled down (Fig 5-6B, Lane 1) The advantage of using the RNA pull-down assay over Northwestern blot is that both C protein and RNA are free to associate or dissociate in solution thus mimicking interaction as in a cellular environment However the lack of

phosphorylation seen on wild type myc-C protein that was pulled down could be due to two reasons (Fig 5-6B, Lane 1, bottom panel) It could be because there exists a population of unphosphorylayted myc-C protein in the cell lysate that was pulled down

Or it could be that there was not enough myc-C protein being pulled down to give a phosphorylation signal on the Western blot (Fig 5-6B, Lane 1, bottom panel)

Nonetheless these assays demonstrate that phosphorylation attenuates RNA binding In addition, they also confirm the functional redundancy found in WNV C

protein (Kofler et al., 2002; Patkar et al., 2007; Schlick et al., 2009) with regards to the

attenuation of RNA binding (Fig 5-6A, Lane 2) Phosphorylation at either terminal of the

C protein only slightly attenuated the RNA binding ability of the C protein and hence this

is consistent with findings that assembly of infectious virions could still occur even when

large sections of the either the amino- or carboxyl- terminal were deleted (Patkar et al., 2007; Schlick et al., 2009)

A straightforward explanation as to why phosphorylation of C protein would

attenuate its affinity for viral RNA in WNV and also other viruses (Gazina et al., 2000; Ivanov et al., 2001; Law et al., 2003) is that phosphates are negatively-charged

Phosphorylation would thus reduce the positive charge on C protein and reduce its affinity for negatively-charged nucleic acids However, phosphorylation does not always

lead to attenuation of nucleic acid binding In some cases there were no effect (Maroto et al., 2000) and in other cases phosphorylation enhanced protein binding to nucleic acid (Green et al., 1992) In these instances, phosphorylation of protein caused a

conformational change which exposed the nucleic acid binding sites on the protein for

interaction with RNA/DNA (Green et al., 1992) Although physical steric changes could

arise from phosphorylation in WNV C protein, it is unlikely that the attenuation of RNA binding observed with WNV C protein was due to conformational changes This is because phosphorylation of WNV C peptides, which are not known to have secondary structures, also attenuated its RNA binding capacity (Fig 4-10) Hence this data supports the idea that WNV C protein and RNA interaction is due to electrostatic forces rather than physical steric forces

However, the enhanced affinity for viral RNA and inefficiency of nuclear localisation of hypophosphorylated C protein was insufficient to explain the lag in virus replication observed in mutant viruses (Fig 6-3) Although complementation with wild type myc-C protein partially restored the virus yield it did not abolish the lag in virus

replication (Fig 6-7) Since in vitro results suggested that dephosphorylation favours the

formation of nucleocapsid (enhanced RNA binding and cytoplasmic localisation of C protein), the ability to phosphorylate C protein must be crucial for the early events of viral replication

In order to circumvent the early events of virus replication, infectious RNA was transfected into the cells and asked if the lag in virus replication could be abolish Indeed, when infectious viral RNA was used to produce mutant virus, the lag observed earlier was abolished (Fig 6-8) At a molecular level, differences between infection and transfection were apparent (Fig 6-10 and Fig 6-11) and consistent with the growth kinetics data RNA and protein synthesis of the mutant virus was slower than the wild type virus (Fig 6-10A and Fig 6-11A) and this corresponded with the lag in virus replication (Fig 6-3) This difference in RNA and protein synthesis between the wild type and mutant virus was also abolished when infectious RNA was transfected into the

cells (Fig 6-10B and Fig 6-11B) This indicated that hypophosphorylation of the C protein did not fundamentally alter the rate of RNA synthesis or protein synthesis, rather the apparent reduction in the rate of RNA and protein synthesis in infected cells is due to some defect caused by the inability of the C protein to phosphorylate

All these data suggested that a C protein amenable to phosphorylation is important for virus replication These results are reminiscent of those observed with

Rubella virus (Law et al., 2003) The authors also discovered a lag in Rubella virus with

hypophosphorylated C protein in infected cells and this lag was overcame by transfection

of infectious Rubella virus RNA They proposed that phosphorylation during the early phase of virus infection was essential for the destabilisation of the C protein to release the viral genome into the cytoplasm for translation and replication Experiments with Potato virus X also suggested that phosphorylation may destabilise C protein since encapsidated viral RNA was made available for translation when its coat protein (analogous with the C

protein) was phosphorylated in vitro (Atabekov et al., 2001) This could also be the

explanation for the observed inhibition of viral RNA and protein synthesis in mutant WNV infection (Fig 6-10A and 6-11A) The high affinity of hypophosphorylated C protein to viral RNA may be inefficient in dissociating encapsidated RNA from the C protein hence causing the viral RNA less available for translation and replication

Whether phosphorylation destabilizes oligomers of WNV C is unknown but observation with hypophosphorylated and wild type myc-C proteins showed that phosphorylation did retard oligomerization (Fig 5-9) It is therefore tempting to suggest that phosphorylation of C protein may also play a role in destabilizing C oligomers during the early phase to aid the release of viral RNA (Fig 7-2) The correlation between

oligomerization and phosphorylation was only demonstrated with the C protein of the HCV virus The authors were however, unable to conclude if oligomerization was needed

for phosphorylation or that oligomerization induced phosphorylation (Shih et al., 1995)

Based on more recent data and observations made in this study with WNV C protein, phosphorylation was not required for the oligomerization of the C protein (Fig 5-9, Lanes 3 - 5, 7 - 9, 13 - 15 and 17 - 19) Nucleocapsid assembly of some viruses in

the Flaviviridae family was achieved in vitro with the addition of nucleic acid Rivero et al., 2004; Kiermayr et al., 2004; Kunkel et al., 2001; Lopez et al., 2009) In

(Acosta-addition, it was reported that the building blocks of the nucleocapsid are dimers of the C

protein (Jones et al., 2003; Kiermayr et al., 2004; Patkar et al., 2007) For WNV C

protein, a crucial residue (Tryptophan 69) required for C protein dimerization has been identified (Bhuvanakantham & Ng, 2005) All these suggested that all the oligomerization information is found on the C protein itself and this process can be spontaneous

Indeed spontaneous dimerization and oligomerization was observed with both the His-C and myc-C proteins (Fig 5-1, Lane 8 and Fig 5-9) Once these oligomers form they do not dissociate to monomers even when boiled in the presence of detergents like SDS This showed that the dimers and oligomers were very stable and resistant to boiling and SDS Since oligomerization is spontaneous, it would be reasonable to suggest that phosphorylation functions to retard oligomerization (Fig 5-9) or destabilisation during the early phase of infection Dephosphorylation of C protein during the late phase of infection on the other hand would hence allow for spontaneous oligomerization of C protein to occur, thus supporting nucleocapsid assembly

In conclusion, this study established various platforms to study the mechanisms of

C protein and RNA interaction It was shown that the C protein does indeed function as a phospho-protein and that phosphorylation and dephosphorylation plays a critical role in regulating the primary function of the C protein, which is the assembly of the nucleocapsid Having shown how phosphorylation modulates the processes of nucleocapsid assembly, a model of how phosphorylation regulates nucleocapsid assembly can be generated from the experiments performed (Fig 7-2) In this model, it is essential for the virus to release the RNA into the cytoplasm for replication during the early phase

of infection and target C protein into the nucleus To do this, not only does the RNA binding affinity of C protein needs to be attenuated but C protein oligomers need to be destabilized to release the viral RNA (Fig 7-2, early phase) It is proposed that during virus entry into the cell, C protein is phosphorylated to destabilize the oligomers and its positive charge attenuated to allow the release of viral RNA into the cytoplasm At the same time, phosphorylation of C protein enhances its interaction with importin-α hence allowing the protein to translocate into the nucleus The significance of the translocation

of C protein into the nucleus serves two purposes One, nuclear localisation of C protein may be needed to effect changes in the host cell to create an environment that favours viral replication, for example, cell cycle arrest Two, shuttling of C protein into the nucleus during the early phase of infection (up to 12 hr post-infection) would prevent its unspecific interaction with viral RNA in the cytoplasm Early or premature packaging may result in encapsidation of aberrant amount of negative strand RNA into the virions

In addition, because C protein is phosphorylated, its affinity for nucleic acid would be attenuated hence further preventing premature associaton with viral RNA At the same

time, phosphorylation of C protein would also retard its oligomerization During the late phase of infection (from 12 hr post-infection onwards), newly synthesized C protein would be dephosphorylated and remain in the cytoplasm for assembly (Fig 7-2, late phase) Dephosphorylation would allow for C protein and RNA interaction and rapid oligomerization for the formation of the nucleocapsid To further prove the validity of this model, future experiments could investigate how C protein is dephosphorylated before packaging Although it was observed tha C protein is dephosphorylated during the course of an infection (Fig 5-10), it is not known which kinase is involved or when is the putative kinase activated to dephosphorylate the C protein Nonetheless, experiments from this study is in agreement with the hypothesis that C protein functions as a phospho-protein and that the dynamic process of phosphorylation and deposphorylation of C protein serves to modulate the processes of nucleocapsid assembly

Figure 7-2 Model of how phosphorylation modulates the functions of C protein to bring

about the assembly of the virus The diagram is divided into two general phases of the virus replication – early phase and late phase The early phase is anytime up to 12 hr post-infection and the late phase is anytime after 12 hr post-infection The dotted red arrow represents the translocation of C protein from into or from the nucleus

REFERENCES

Acosta-Rivero, N., Rodriguez, A., Musacchio, A., Falcon, V., Suarez, V M.,

Martinez, G., Guerra, I., Paz-Lago, D., Morera, Y., de la Rosa, M C., Morales-Grillo, J & Duenas-Carrera, S (2004) In vitro assembly into virus-

like particles is an intrinsic quality of Pichia pastoris derived HCV core protein

Biochem Biophys Res Commun 325, 68-74

Ada, G L & Anderson, S G (1959) Yield of infective ribonucleic acid from impure

Murray Valley encephalitis virus after different treatments Nature 183, 799-800

Anderson, J F., Andreadis, T G., Vossbrinck, C R., Tirrell, S., Wakem, E M.,

French, R A., Garmendia, A E & Van Kruiningen, H J (1999) Isolation of

West Nile virus from mosquitoes, crows, and a Cooper's hawk in Connecticut

Science 286, 2331-2333

Atabekov, J G., Rodionova, N P., Karpova, O V., Kozlovsky, S V., Novikov, V K

& Arkhipenko, M V (2001) Translational activation of encapsidated potato

virus X RNA by coat protein phosphorylation Virology 286, 466-474

Banks, J D & Linial, M L (2000) Secondary structure analysis of a minimal avian

leukosis-sarcoma virus packaging signal J Virol 74, 456-464

Beasley, B E & Hu, W S (2002) cis-Acting elements important for retroviral RNA

packaging specificity J Virol 76, 4950-4960

Bernard, K A & Kramer, L D (2001) West Nile virus activity in the United States,

2001 Viral Immunol 14, 319-338

Berthet, F X., Zeller, H G., Drouet, M T., Rauzier, J., Digoutte, J P & Deubel, V

(1997) Extensive nucleotide changes and deletions within the envelope

glycoprotein gene of Euro-African West Nile viruses J Gen Virol 78 ( Pt 9),

2293-2297

Bhuvanakantham, R., Cheong, Y K & Ng, M L (2010) West Nile virus capsid

protein interaction with importin and HDM2 protein is regulated by protein kinase

C-mediated phosphorylation Microbes Infect 12, 615-625

Bhuvanakantham, R., Chong, M K & Ng, M L (2009) Specific interaction of capsid

protein and importin-alpha/beta influences West Nile virus production Biochem

Biophys Res Commun 389, 63-69

Bhuvanakantham, R & Ng, M L (2005) Analysis of self-association of West Nile

virus capsid protein and the crucial role played by Trp 69 in homodimerization

Biochem Biophys Res Commun 329, 246-255

Bradford, M M (1976) A rapid and sensitive method for the quantitation of microgram

quantities of protein utilizing the principle of protein-dye binding Anal Biochem

72, 248-254

Briese, T., Rambaut, A., Pathmajeyan, M., Bishara, J., Weinberger, M., Pitlik, S &

Lipkin, W I (2002) Phylogenetic analysis of a human isolate from the 2000

Israel West Nile virus epidemic Emerg Infect Dis 8, 528-531

Brinton, M A (2002) The molecular biology of West Nile Virus: a new invader of the

western hemisphere Annu Rev Microbiol 56, 371-402

Brinton, M A., Fernandez, A V & Dispoto, J H (1986) The 3'-nucleotides of

flavivirus genomic RNA form a conserved secondary structure Virology 153,

113-121