Characterisation of a Human DEAD-Box Protein (DDX3) and Its Interaction With Hepatitis C Virus Core Protein

Hepatitis C Virus (HCV) core protein is believed to form the viral nucleocapsid. However, numerous reports suggest it can also modulate diverse cellular processes. It is possible that at least some of these pleiotropic effects are exerted through the interaction of core protein with a range of host cellular factors, including a putative RNA helicase of the DEAD-box family termed DDX3. The main aims of this study were to i) characterise DDX3, in terms of its basic properties and normal role in cellular metabolism, and ii) investigate the interaction of DDX3 with core protein and determine any influence of this association on HCV replication/pathogenesis. A number of anti-DDX3 immunological reagents were already available for study of the endogenous DDX3 protein, as well as various truncated or mutated forms of the protein that were subsequently cloned and expressed in a variety of systems. Core protein was produced using recombinant vaccinia virus (rVV) due to the lack of an efficient cell culture system for HCV. To allow comparisons with natural infection of permissive cells with this hepatotropic virus, studies were usually limited to human hepatocyte-derived cell lines, while core protein was generally expressed along with the HCV glycoproteins (E1-E2), as it would be in vivo, to ensure proper processing of core. Recombinant baculoviruses (rbacs) carrying the DDX3- and core-coding sequences were generated for further examination of these proteins. Since little was generally known about DDX3, initial studies concentrated on its fundamental characteristics, including investigations into expression of its mRNA transcript and protein in human hepatocytes and other mammalian cell lines. The DDX3 mRNA transcript was also studied in a wide range of human tissues. These analyses strongly suggest that DDX3 is a ubiquitous and highly conserved cellular protein. Consistent with previous reports regarding the DDX3/core interaction, expression of core protein in hepatocytes led to a marked redistribution of endogenous or over-expressed DDX3. This redistribution of DDX3 in the presence of core also occurred in the recently described HCV sub-genomic replicon- expressing cell lines. These observations indicate that core protein aberrantly sequesters a ubiquitous, highly conserved cellular protein, likely disrupting its potentially crucial function. Intriguingly, further studies suggested that core protein directly or indirectly modifies DDX3. An anti-DDX3 polyclonal antibody (PAb) specifically detected DDX3 in insect cell extracts previously infected with rbac expressing the protein, and detected the endogenous DDX3 in human hepatocytes; co-expression of rbacs expressing core (or core-El-E2) and DDX3 in insect cells, or infection of human hepatocytes with rVV expressing core-El-E2 led to the appearance of a higher molecular weight isoform of DDX3. This provides further evidence that the DDX3/core interaction is genuine, and possibly emphasises its significance in terms of HCV pathogenesis. Several insights into DDX3 and its interaction with core protein were given by expression of DDX3 mutants from mammalian expression plasmids. Of particular interest was a mutant containing a single amino acid change within the DEAD-box, a motif that is highly conserved amongst members of the large family of known and putative RNA helicases to which DDX3 belongs. This mutant showed a very distinct subcellular distribution compared with the wild-type protein, although it retained its ability to interact with core. In collaboration with others, it was shown that this DDX3 mutant was enzymatically incapacitated, consistent with the involvement of the DEAD-box in ATP hydrolysis. These data suggest important features regarding DDX3 and its interaction with core: i) the functional capabilities of DDX3 are linked to its subcellular localisation; ii) the normal distribution of DDX3 is irrelevant for its association with core, possibly indicating that their interaction occurs prior to subcellular targeting of DDX3; iii) the enzymatic competence of DDX3 is not essential for its interaction with core. A putative nuclear export signal (NES) was also identified in DDX3 by comparison with its Xenopus laevis homologue. ANES-DDX3, lacking the N-terminal 21 amino acids of the protein, was cloned and expressed by plasmid in hepatocytes as before. However, although this protein appeared to be more concentrated in the nuclear periplasm, accumulation of the protein within the nucleus itself was not detected. This could suggest that the putative NES of DDX3 is not functional in vivo, or that more than one mechanism governs its nucleocytoplasmic transport. Consistent with the latter hypothesis, subcellular fractionation of hepatocyte cell extracts revealed a small quantity of DDX3 protein in the nucleus