Mechanism of Transcription and Replication

The enzymology of HDV RNA replication has commanded considerable interest over the last few years. Host cell RNA polymerase II, which normally uses a DNA template, has long been implicated as the likely replicase for HDV (Macnaughton et al. 1991). Recent studies based on a-amanitin resistance (see above) have confirmed the major role this polymerase is likely to play (Macnaughton et al. 2002; Chang and Taylor, 2002). It is also clear that

RNA polymerase II is likely responsible for transcription of the mRNA for HDAg (Modahl et al. 2000). In contrast, it seems likely that a different enzyme (probably pol I or a pol I-like enzyme) is responsible for the sysnthesis of full-length antigenomic HDV RNA (Modahl et al. 2000; Macnaughton et al. 2002). However, this view is still somewhat controversial. Recently, a report from the Taylor laboratory (Nie et al. 2004), using various HDV cDNA constructs under the control of an SV40 promoter, concluded that poly(A) addition or ribozyme cleavage were alternative processing outcomes to generate the separate antigenomic HDV RNA transcripts. Significantly, both processing events occurred on pol II-derived transcripts. This led the authors to conclude that only one polymerase (pol II) is required for HDV RNA replication. The system used in this study was designed to be replication-defective and thus overall bears little resemblance to natural HDV RNA-templated RNA replication. Moreover, a disproportionally high molar ratio of mRNA species to full-length antigenomes (3:1) was obtained as compared to that observed in HDV-infected livers (1:50), indicating that the system used for this study was prone to artifacts. In contrast, evidence for the multiple polymerase model is based primarily on RNA-only transfection studies, which more closely represents the natural situation. Moreover, using BrUTP labeling, we have recently observed that de novo genomic RNA synthesis was sensitive to a-amanitin and occurred in the vicinity of PML bodies whereas antigenomic synthesis was a-amanitin resistant and occurred either in or at the periphery of the nucleolus (Li et al. 2006).

The involvement of different transcription machineries for genomic and antigenomic RNA synthesis is likely crucial for the HDV life cycle. For example, genomic RNA must be exported to the cytoplasm for packaging. Since pol II-mediated transcription is coupled to the nuclear export machinery (reviewed by Cullen 2003), this presents a convenient method for the export of genomic HDV RNA immediately following its synthesis. Interestingly, while the cellular mRNA export event is linked to splicing, only the completely or nearly completely processed forms of genomic HDV RNA (predominantly monomers) are exported. Thus, it is tempting to speculate that in the case of HDV, the ribozyme cleavage event, which is somewhat analogous to splicing, is linked to export. In contrast, the potential involvement of pol I in synthesis of antigenomic HDV RNA maybe essential for the production of the full-length species. Specifically, pol I transcripts are never polyadenylated and production of full-length antigenomes of HDV requires that the polyadenylation signals (used during HDAg mRNA production) be silenced.

How then can RNA pol II and other cellular polymerases use an RNA template when the normal template is DNA? There are several possibilities for this, none of which are mutually exclusive.

1. HDAg shares some properties with transcription factors. Thus, direct or indirect binding of HDAg with the host polymerase (and/or other transcription factors) may lead to a relaxation of normal template requirements. Indeed, HDAg has been shown to bind to both HDV RNA and pol II (Lin et al. 1990; Chao et al. 1991; Yamaguchi et al. 2001).

2. The rod-shaped structure of HDV RNA resembles double-stranded DNA; therefore, it is conceivable that cellular RNA polymerases and transcription factors may recognize double-stranded RNA. In this regard, it has been shown that circular monomeric HDV cDNA contains endogenous promoters capable of directing HDV RNA synthesis (Macnaughton et al. 1993b). Interestingly, one of these, located near the transcription initiation site for HDAg mRNA, also exhibited promoter activity as an RNA molecule (Beard et al.; see Fig. 1: putative RNA promoter)

The high GC content and extensive secondary structure would suggest that HDV RNA is a challenging template. Not surprisingly then, one of the functions attributed to S-HDAg is the promotion of elongation bypol II (Yamaguchi et al. 2001). Nevertheless, it is possible that RNA synthesis proceeds in a somewhat stop-start fashion with frequent polymerase pauses at regions of high secondary structure. In extreme cases, the polymerase may actually detach and subsequently re-anneal. Evidence for the latter comes from the likely intermolecular template switching required to reconstitute replicating HDV RNA from pairs of less than full-length RNAs (Gudima et al. 2005). Moreover, RNA recombination, which depends on template switching, has been demonstrated for HDV (Wang and Chao 2005). Examination of the fine detail of the RNA secondary structure of the HDV rod (Fig. 1) reveals that the protein-coding domain consists of stretches of relatively intense intramolecular base-pairing 30-40 nucleotides in length, interspaced with single-stranded regions of 5-12 bases. It is tempting to speculate that these single-stranded domains help the polymerase move along the template by providing regions that can be easily dissociated. In contrast, the most highly intramolecular base-paired region of HDV RNA occurs in the viroid-like domain (Fig. 1). Significantly, for antigenomic HDV RNA synthesis, this region lies immediately downstream of the polyadenylation signal. Thus, polymerase pausing in this region may assist in the exonuclease-dependent termination of transcription demonstrated for cellular mRNA production (Kim et al. 2004; West et al. 2004).

Finally, it remains a possibility that HDV RNA replication involves a cytoplasmic phase. While this is required for packaging of the genomic RNA into virions, it is curious that RNA export happens well before the virus assembly can take place. Moreover, HDV RNA has been shown to continually shuttle between the nucleus and cytoplasm (Tavanez et al. 2002). Thus, it is conceiv able that certain steps of HDV replication cycle, e.g. RNA ligation, take place in the cytoplasm. Thus, HDV RNA shuttling between the cytoplasm and the nucleus may be a critical step for successful RNA replication.

0 0

Post a comment