Introduction

Top Abstract Introduction Results and Discussion Materials and methods References

It is thought that the majority of eukaryotic mRNAs initiate translation by a scanning mechanism, in which 40S ribosome subunits are assembled at the 5' cap by translation initiation factors, which then promote a 5' to 3' nucleotide-by-nucleotide search for the initiating AUG codon (Jackson 1998). Some mRNAs initiate translation internally in the absence of a cap-dependent scanning process, which is directed by a poorly understood internal ribosome entry site (IRES; Jackson 1998). A small group of mRNAs identified to date, initiate translation by ribosome shunting, which combines features of both scanning and internal initiation (Jackson 1996). In shunting, ribosomes are loaded onto mRNA by a cap-dependent process but then shunt or bypass large segments of the mRNA before initiating translation at a downstream AUG (Jackson 1996). Ribosome shunting has been described for Cauliflower mosaic virus (CaMV) 35S mRNA (Futterer et al. 1993), Sendai virus Y protein mRNA (Curran and Kolakofsky 1988; Latorre et al. 1998), the family of adenovirus late mRNAs (Yueh and Schneider 1996), and papillomavirus E1 mRNA (Remm et al. 1999).

There are likely two functions for ribosome shunting. First, as described for CaMV 35S mRNA (Futterer et al. 1993), shunting can expand the coding capacity of an mRNA by permitting ribosomes to bypass upstream ORFs, directing ribosomes to an internal coding region that would not ordinarily be translated. Second, as described for adenovirus late mRNAs (Yueh and Schneider 1996), shunting can confer selective translation, despite inhibition of cell protein synthesis. For instance, during adenovirus late infection, or heat shock of cells, translation of most capped mRNAs is inhibited, and this inhibition is associated with dephosphorylation of translation initiation factor eIF4E (Huang and Schneider 1991; Zhang et al. 1994; Duncan 1996; Feigenblum and Schneider 1996; Schneider 1996). eIF4E is a component of the cap-binding initiation factor complex (eIF4F), which includes eIF4A, an ATP-dependent RNA helicase, and eIF4G, a large adapter protein upon which the complex assembles at the cap (Hentze 1997). Phosphorylation of eIF4E correlates positively with increased stability of the complex on the capped mRNA, increased cap-dependent RNA unwinding activity, and, therefore, increased ribosome loading onto mRNA (Rozen et al. 1990; Haghighat et al. 1995). Thus, when eIF4F-cap association is impaired, eukaryotic mRNAs that translate by either internal or shunting processes continue to initiate, probably because little or no eIF4F unwinding activity is required (Yueh and Schneider 1996).

The mechanism of ribosome shunting has not been described in molecular detail. CaMV 35S and late adenovirus mRNAs can switch between scanning and shunting mechanisms of initiation, which are thought to be mutually exclusive due to disruption of RNA shunting elements by scanning ribosomes (Yueh and Schneider 1996; Hohn et al. 1998). In CaMV, a short ORF (sORF) followed by one or more stable hairpin structures in the 5' noncoding region (NCR) comprises a minimal shunting element (Hemmings-Mieszczak et al. 1997, 1998; Dominguez et al. 1998; Pooggin et al. 1998). This configuration is thought to provide 40S ribosome subunits to the shunting elements for nonlinear downstream translation. Ribosome shunting is directed in late adenovirus mRNAs by the tripartite leader, a highly conserved 200-nucleotide 5' NCR found on most late adenovirus mRNAs (Yueh and Schneider 1996). The tripartite leader is thought to contain a 25- to 44-nucleotide unstructured 5' conformation, followed by a complex group of stable hairpin structures (Zhang et al. 1989; Dolph et al. 1990; Yueh and Schneider 1996), but no sORF. The internal hairpins are essential components of the shunting element (Yueh and Schneider 1996). The tripartite leader provides preferential translation to late adenovirus mRNAs after viral inhibition of cell protein synthesis (Huang and Schneider 1990; Logan and Shenk 1984) in heat-shocked cells and in growth-arrested cells (Huang and Schneider 1990; Feigenblum and Schneider 1996; Yueh and Schneider 1996), which have been shown to have impaired eIF4F activity, possibly resulting from dephosphorylation of eIF4E. When phosphorylated eIF4E is abundant, and, therefore, cap-dependent eIF4F complex activity is high, the tripartite leader directs translation initiation by both 5' scanning and shunting mechanisms. When eIF4E is dephosphorylated, the tripartite leader exclusively and efficiently directs initiation by shunting (Yueh and Schneider 1996).

We have investigated the molecular mechanism by which the tripartite leader directs efficient initiation by ribosome shunting. We show that the tripartite leader promotes a high level of ribosome shunting through the use of three conserved sequences that are complementary to the stem of the 3' hairpin of 18S rRNA. Ribosome shunting was also found to be directed by the adenovirus IVa2 5' NCR, which is also translated during the late phase of infection, and contains a related comlementarity to 18S rRNA. Moreover, the human heat shock protein 70 (hsp70) 5' NCR contains one element related to the tripartite leader 18S rRNA complementarity, which is shown to promote ribosome shunting on this mRNA during heat shock, when eIF4F-cap-dependent protein synthesis is blocked. These results highlight the novel role of complementarity to 18S rRNA in some forms of ribosome shunting.

	Results and Discussion

Top Abstract Introduction Results and Discussion Materials and methods References

Ribosome shunting facilitated by adenovirus tripartite leader complementarity to 18S rRNA

Ribosome shunting directed by the adenovirus late tripartite leader occurs with high efficiency at roughly similar levels to that of scanning initiation (Yueh and Schneider 1996). This efficiency contrasts with an ~10-fold lower level of shunting found in CaMV, Sendai virus, and papillomavirus mRNAs (Futterer et al. 1993; Latorre et al. 1998; Remm et al. 1999). A feature present in the tripartite leader that is not found in these other mRNAs is a group of three elements of split (bipartite) complementary to both RNA strands of the stem in the 3' hairpin of 18S rRNA (Fig. 1A). 18S rRNA is located in the 40S (small) ribosome subunit. There is no known role for the 18S rRNA 3' hairpin in translation although it is highly conserved in prokaryotes and eukaryotes and is located on the outside surface of the 40S subunit (Green and Noller 1997) where it is accessible for interactions with other RNAs (Moore 1996). Tripartite leader complementarity to the 18S rRNA 3' hairpin is conserved in human and mammalian viral serotypes (Dolph et al. 1990; Yueh and Schneider 1996).

View larger version (51K): [in this window] [in a new window]   Figure 1.   Scanning and shunting translation of tripartite leader mutants. (A) (Left) Diagram showing the 3' hairpin of 18S rRNA and regions of the stem complementary to adenovirus tripartite leader elements (solid rectangles on stem). (Right) Bipartite sequence complementarity of the 3' hairpin of human 18S rRNA and the human Ad5 tripartite leader. (C1-C3) Three complementarities in the tripartite leader (boxed), with nucleotide positions indicated in parentheses relative to the +1 start site of tripartite leader transcription. (B) Wild-type and deletion mutants of the tripartite leader are shown, ± insertion of a stable hairpin structure (B202) in the 3' end of the NCR to block scanning-dependent initiation. Deletions are indicated by solid rectangles, with the position of junction sites indicated. (C) Steady-state Northern analysis of HBsAg reporter mRNAs obtained from equal numbers of transfected HeLa cells. Blots were hybridized to a 32P-labeled probe prepared from the HBsAg coding region. (D) ELISA analysis of HBsAg levels obtained from five independent transfections of HeLa cells performed at similar efficiencies with the different tripartite leader 5' NCRs. Transfection efficiencies (30%-35%) were determined by cotransfection of pGFP. Data were normalized relative to the average translational activity of the wild-type tripartite leader (WT). Mutant 3LDR contains a 50-nucleotide (BamHI-SalI) fragment of pBluescriptII SK polylinker in place of the tripartite leader.

We carried out mutational analyses of the tripartite leader to identify critical RNA elements that facilitate a high level of ribosome shunting. Specific deletions were introduced into a cDNA copy of the tripartite leader by oligonucleotide-directed PCR-mediated mutagenesis. The position of tripartite leader complementarities to 18S rRNA (C1, C2, and C3) are shown, as are the boundaries of the nucleotide segments deleted within a cDNA copy of the 220-nucleotide tripartite leader (Fig. 1B). To determine its ability to direct scanning and shunting initiation, wild-type and mutant forms of tripartite leader cDNAs were fused to the coding region of the hepatitis B virus surface antigen (HBsAg) gene as a reporter. The level of HBsAg secreted into the medium was then precisely quantitated by ELISA. Two sets of 5' NCRs were constructed, with and without a stable hairpin at the 3' end of the tripartite leader. Insertion of the stable hairpin at this position was shown previously to block scanning but not shunting ribosome initiation, providing a direct measure of the level of shunting translation (Yueh and Schneider 1996). HeLa cells were transiently transfected at similar efficiencies with plasmid reporter constructs, as determined by inclusion of a green fluorescent protein expression vector (data not shown). Equivalent steady-state levels of HBsAg reporter mRNAs were observed by Northern analysis (Fig. 1C), and translation of HBsAg mRNAs was quantified by ELISA (Fig. 1D).

As found previously, ribosome shunting accounts for 50%-60% of the normal translation activity of the wild-type tripartite leader (3LDR), shown by insertion of a stable 3' hairpin (mutant B202; Fig. 1D). Translation of mRNAs was normalized to the unmodified wild-type tripartite leader (set to 100% activity). Replacement of the leader with a 50-nucleotide foreign RNA segment fully blocked shunting translation initiation, while reducing scanning initiation by only ~50% (3LDR samples). Deletion of the majority of the tripartite leader abolished ribosome shunting [mutants D8 and D9 (±B202)] but only slightly reduced its ability to direct scanning initiation (two- to threefold). Collectively, these results confirm an essential requirement for the tripartite leader in ribosome shunting. Tripartite leader complementarities C1-C3 were then mutated individually, or in combination, to assess their impact on scanning and shunting initiation. Individual deletion of complementarities C1, C2, or C3 (mutants D10, D12, and D13, respectively) had a slight negative impact on the overall initiation of translation. Deletion of only C1 had no effect on shunting initiation (D10/B202). Deletion of only C2 (D12/B202), or of most of C3 (D13/B202), reduced shunting initiation only slightly, by ~25% compared with B202. Similarly, small deletions of the tripartite leader outside the C1-C3 complementarity regions did not diminish scanning or shunting initiation of translation (mutants D4 and D11). These results implicate involvement of C2 or C3 complementarities in ribosome shunting. Therefore, it was determined whether complementarities C1-C3 possess functional redundancy that might mask the effects of individual deletion. Paired deletions of C1 and C2 (mutant D14/B202) reduced shunting slightly, to 60% of B202 translation activity, and had little if any negative effect on overall (scanning) initiation. Paired deletion of C1 and C3 (mutant D15/B202) reduced shunting initiation threefold to 20% of B202 translation activity. However, combined mutation of C2 and C3 (mutant D16/B202) strongly impaired shunting (by ~20-fold) to ~5% of B202 translation activity. The moderate decrease in overall translation activity of mutant D16 (by ~20%) is probably accounted for solely by the loss of shunting translation. Therefore, these results demonstrate that complementarities C1-C3 are specifically involved in facilitating a high level of translation initiation by ribosome shunting. They also indicate that there is redundancy and a hierarchy in the importance of these elements, in which C2 and C3 are functionally dominant over C1. Collective mutation of C1-C3 (mutant DL3) did not further reduce ribosome shunting or scanning initiation beyond that of the C2-C3 mutant (D16 and D16/B202). Paired mutation of similar-sized segments outside of C1, C2, or C3 (mutants D18 and D19) did not detectably reduce scanning or shunting initiation. Finally, these data also demonstrate that the tripartite leader contains a basal shunting element that promotes ~5% shunting activity in the absence of specific complementarity to 18S rRNA. The basal element remains to be identified, but it is evident that it is strongly augmented by a specific complementarity to the 3' hairpin of 18S rRNA.

Next, we examined the importance of tripartite leader complementarities in directing ribosome shunting during late adenovirus infection, when eIF4F-dependent cell protein synthesis is inhibited. 293 cells, which are sensitive to shutoff of protein synthesis by adenovirus, were transfected with plasmids expressing HBsAg reporters linked to the wild-type tripartite leader, or mutant leaders, then infected 18 hr later with wild-type adenovirus. The level of reporter translation was determined prior to shutoff of cell protein synthesis (from 10 to 16 hr) or after shutoff (from 24 to 30 hr), by measurement of the amount of HBsAg secreted into the medium during this 6-hr period (Table 1). Inhibition of host cell protein synthesis and dephosphorylation of eIF4E at 24 hr post-infection was confirmed as described (Yueh and Schneider 1996), and Northern analysis demonstrated only a 25% decrease in cytoplasmic levels of plasmid transcribed mRNAs during the late phase of virus infection (data not shown). Translation of the wild-type tripartite leader (3LDR) and a tripartite leader containing the 3' B202 hairpin (B202) increased (30%-40%) during late adenovirus infection, after inhibition of cell protein synthesis as observed previously (Yueh and Schneider 1996). The increase in shunting is thought to result from reduced disruption of the RNA shunting elements by scanning ribosomes (Yueh and Schneider 1996). Specific deletion of tripartite leader complementarity to 18S rRNA also impaired shunting translation during late virus infection. Mutants D12 and D13, which lack elements C2 or C3, respectively, translated at levels similar to wild-type during early infection. Mutant D14, which lacks regions C1 and C2, and mutant D15, which lacks regions C1 and C3, also translated normally during early adenovirus infection. However, during late adenovirus infection, mutants D12 and D13 translated at about half the level of the B202 control mRNA, and mutant D15 translated at only ~30% the level of B202. Mutant D16, which is deleted of complementarities C2 and C3, was only efficiently translated prior to viral inhibition of cell translation and was poorly translated (~5%) during late infection, following shutoff of host protein synthesis. These results are consistent with a hierarchy in functional importance of complementarities C2 and C3 in ribosome shunting, as determined in Figure 1 by insertion of secondary structure within the tripartite leader. These data also show that complementarity to 18S rRNA is vital for efficient tripartite leader translation by ribosome shunting during late adenovirus infection, when eIF4F activity and host cell protein synthesis are impaired.

Adenovirus IVa2 mRNA is translated by shunting that is dependent on complementarity to the 18S rRNA 3' hairpin

Adenovirus late IVa2 mRNA is synthesized during the intermediate phase of virus infection. It lacks the tripartite leader 5' NCR, but nonetheless is well translated during late viral infection after inhibition of eIF4F-dependent protein synthesis (Hayes et al. 1990; Huang and Schneider 1990). Furthermore, the adenovirus late IX intermediate phase mRNA, which also lacks the tripartite leader, is weakly translated during the late phase of infection (Hayes et al. 1990; Huang and Schneider 1990). Examination of these mRNAs revealed that the IVa2 5' NCR sequence contains two regions strongly complementary to the 18S rRNA 3' hairpin, which are similar to sequences found in the tripartite leader (Fig. 2A). No significant complementarity to 18S rRNA was found in the IX mRNA 5' NCR. cDNA copies of the IVa2 and IX 5' NCRs were generated by synthetic oligonucleotide-mediated reverse transcription of late adenovirus mRNA, and the DNA sequence was verified. 5' NCRs were tested for ribosome shunting activity by fusion to the HBsAg coding region, with and without insertion of a stable hairpin structure in the 3' end of the NCR (Fig. 2B). The IVa2 5' NCR directed strong ribosome shunting activity, averaging ~30% the level of the tripartite leader (cf. B202 and IVa2/B202). Deletion of the majority of the IVa2 5' NCR, including a region containing the two complementarities to the 18S rRNA 3' hairpin (mutant IVa2) specifically abolished shunting. A mutant deleted in the majority of the two IVa2 complementarities (mIVa2, Fig. 2B) directed ribosome shunting very poorly (~2% of that of wild type). This mutant also displayed near normal overall translation activity, which was reduced by only ~30%, likely reflecting the loss of shunting translation. These data demonstrate that the IVa2 5' NCR directs translation initiation by both ribosome shunting and scanning. They also specifically implicate complementarity to the 3' hairpin of 18S rRNA in strong shunting activity.

View larger version (34K): [in this window] [in a new window]   Figure 2.   Scanning and shunting translation directed by the Ad5 IVa2 5' NCR. (A) Bipartite sequence complementarity of the 18S rRNA 3' hairpin and the 5' NCR of human Ad5. Complementary elements C1 and C2 (boxed) and nucleotide positions in the IVa2 and 5' NCR are indicated relative to the mRNA transcription start site. (B) Wild-type and deletion mutants of the Ad5 IVa2 and IX 5' NCRs are shown, ± insertion of the B202 stable hairpin structure. The boundary of deletions are shown in cDNA copies of the 5' NCRs are shown above solid rectangles. (C) ELISA analysis of HBsAg reporter mRNAs represent an average of five independent transfections of HeLa cells at similar efficiencies. The IVa2 mutant is deleted from position 1 to position 112 of the IVa2 5' NCR. Steady-state Northern analysis of cytoplasmic reporter HBsAg mRNAs for different constructs was performed with HeLa cells transfected at similar efficiencies.

Next, the ability of the IVa2 5' NCR to direct translation during late adenovirus infection was investigated. 293 cells transfected with reporter expression vectors were later infected with adenovirus, and translation of HBsAg was examined prior to and after inhibition of cell protein synthesis during the late phase of adenovirus infection (Table 1). Prior to adenovirus shutoff of cell protein synthesis, the IVa2 5' NCR directed translation at ~40% the level of the tripartite leader, which was only reduced to 25% during the late phase of infection. Insertion of the B202 hairpin in the 3' end of the IVa2 5' NCR did not further reduce its ability to direct translation during late adenovirus infection (mutant IVa2/B202). These results indicate that the IVa2 mRNA translates at a high level by ribosome shunting during late adenovirus infection, despite inhibition of host cell protein synthesis. The IVa2 mRNA translated during early adenovirus infection at a level similar to that of the wild-type IVa2 mRNA but was reduced to <2% after inhibition of cell protein synthesis during the late phase of infection. The IX mRNA 5' NCR did not translate by ribosome shunting, as it was inhibited by insertion of the B202 hairpin (Fig. 2). When expressed in transfected cells, the IX 5' NCR directed translation only during the early phase of adenovirus infection (Table 1). We propose that the low translation level of the IX mRNA during the late phase of natural adenovirus infection results from the very large abundance of the transcript that is synthesized from the viral genome (Huang and Schneider 1990).

Translation of human hsp70 mRNA by a ribosome shunt

To determine whether a bipartite complementarity to the 18S rRNA 3' hairpin might be involved in facilitating ribosome shunting in other mRNAs, we performed a database search for 5' NCR elements related to the bipartite sequence found in the tripartite leader and IVa2 mRNAs. The human hsp70-1 5' NCR, but not hsp70 from Drosophila melanogaster, contains a single element strongly related to that found in the tripartite leader (Fig. 3A). A full-length cDNA copy of the human and Drosophila hsp70 5' NCRs were generated from commercial HeLa cell and Drosophila cDNA libraries by synthetic oligonucleotide-mediated PCR, and the DNA sequence was verified and then fused to the HBsAg coding region. 5' NCRs were examined for their ability to direct translation by scanning or ribosome shunting, as measured by insertion of a stable hairpin in the distal end of the 5' NCR (Fig. 3B). Both the wild-type human and Drosophila hsp70 5' NCRs directed strong translation relative to the tripartite leader in transiently transfected cells (~80% and 110%, respectively). Scanning-dependent initiation was fully blocked by insertion of a stable hairpin into the 3' end of the Drosophila 5' NCR, which lacks 18S rRNA complementarity. However, the human hsp70 5' NCR continued to direct translation at ~30% the wild-type level, despite insertion of the hairpin. These results indicate that the human hsp70 5' NCR directs translation by both scanning and ribosome shunting.

View larger version (26K): [in this window] [in a new window]   Figure 3.   Translation activity of human hsp70-1 5' NCR. (A) Bipartite complementarity of the human hsp70 5' NCR with the 3' hairpin of 18S rRNA. Complementary regions are boxed and indicated by nucleotide positions relative to the natural start of transcription in the 5' NCR. (B) ELISA analysis of HBsAg reporter mRNAs represented by three independent transfections of HeLa cells at similar efficiencies. Northern analysis was performed on cytoplasmic mRNAs obtained from HeLa cells transfected at similar efficiencies.

Studies were carried out to independently confirm shunting in the human hsp70 5' NCR by measurement the ability to direct translation in transiently transfected cells following heat shock and inhibition of translation. Transfected HeLa cells were subjected to heat shock at 44°C for 4 hr. These conditions block eIF4E phosphorylation and protein synthesis by 1 hr of heat treatment (Yueh and Schneider 1996; Laroia et al. 1999; data not shown). Translation of HBsAg from 1 to 4 hr after heat shock of cells was then determined (Table 2), by measurement of the level of protein secreted into medium during that time. Heat shock had no detectable affect on the cytoplasmic steady-state level of mRNAs (data not shown). Heat shock inhibition of eIF4F-dependent protein synthesis was evident in the translational inhibition of the 3LDR construct, which lacks the tripartite leader, and the inhibition of the Drosophila hsp70/B202 5' NCR. The human hsp70 5' NCR, however, continued to translate strongly following heat shock regardless of whether the B202 stable hairpin was inserted in the 3' end of the NCR. The human hsp70 5' NCR did not function as an IRES when inserted internally in a dicistronic mRNA (data not shown), excluding the possibility of internal ribosome entry. A mutant deleted of the majority of the human hsp70 5' NCR (Huhsp70) was translated at about one-third the level of the wild-type 5' NCR at normal temperature, but failed to translate during heat shock. To determine whether complementarity to 18S rRNA promotes shunting translation on the human hsp70 5' NCR, a mutant was developed that is deleted of region 150-215, which encompasses the 3' 18S rRNA complementarity (Hu mhsp70). Mutant Hu mhsp70 directed efficient translation at normal temperature but was specifically impaired during heat shock, reduced in translation activity by threefold. These results demonstrate that the human hsp70 5' NCR translates by both scanning and ribosome shunting mechanisms, and the shunting mechanism is used exclusively by the hsp70 mRNA during heat shock. They also implicate elements complementary to the 3' hairpin of 18S rRNA and related to those found in the tripartite leader and IVa2 5' NCRs in promoting a stronger level of ribosome shunting in the human hsp70 mRNA.

The molecular mechanism by which complementarity to 18S rRNA promotes ribosome shunting in the 5' NCRs is not known. It is unlikely that these complementary elements singularly comprise the entire RNA shunting element, particularly because detectable ribosome shunting activity is retained after deletion of these elements in the tripartite leader, IVa2, and human hsp70 5' NCRs. In this regard, tripartite leader mutants devoid of C2 and C3 complementarities direct ribosome shunting at a low level (~5%), similar to that of CaMV, Sendai virus, and papillomavirus 5' NCRs (Curran and Kolakofsky 1988; Futterer et al. 1993; Remm et al. 1999), all of which lack a complementarity to the 3' hairpin of 18S rRNA. On the basis of secondary structure mapping of the tripartite leader (Zhang et al. 1989) and structure prediction analysis tripartite leader complementarities C1, C2, and C3 may be partially or predominantly found in unstructured regions and single-stranded loops. This situation is reminiscent of the placement of Shine-Dalgarno elements in prokaryotic mRNAs. For example, bacteriophage MS2 mRNA translation was impaired if complementarity to 16S rRNA was embedded within secondary structure (de Smit and Van Duin 1994). There is no direct evidence, however, that complementarity to the 3' hairpin of 18S rRNA within 5' NCRs actually promotes rRNA-mRNA interaction. Although a direct interaction between 18S rRNA and the tripartite leader is potentially possible as a result of cooperative processes such as transient on/off base-pairing, technically, this interaction will be very difficult to detect at the molecular level. The 5' NCR of S15 ribosomal protein mRNA and the Gtx homeodomain protein mRNA 5' NCR, possess a striking complementarity to the 3' hairpin of 18S rRNA that largely coincides with the C2 and C3 elements of the tripartite leader (Tranque et al. 1998; Hu et al. 1999). In these mRNAs, which initiate translation by ribosome scanning, the complementarity actually impairs translation initiation, possibly because 40S ribosome subunits are tethered to the mRNA (Tranque et al. 1998; Hu et al. 1999). Deletion of the element enhances scanning-mediated initiation. Therefore, it is possible that shunting involves stable association of 40S ribosome subunits to shunting elements containing complementarity to 18S rRNA. Studies now need to determine whether shunting involves direct 18S rRNA-mRNA interaction, or a specific RNA conformation directed by the complementarity to 18S rRNA, and whether binding of 40S subunits or initiation factors are then selectively recruited to the mRNA.

	Materials and methods

Top Abstract Introduction Results and Discussion Materials and methods References

Molecular cloning and site-directed  mutagenesis

Deletions, point mutations, and inversions were introduced in the human Ad5 serotype tripartite leader cDNA in vector pSV-HBV (Dolph et al. 1988, 1990; Yueh and Schneider 1996). Deletions were generated by overlap-extension PCR with complimentary oligonucleotides. The boundaries of retained sequences are indicated in Figure 1 and 2 diagrams. Nucleotide positions are shown relative to the +1 position (transcription start site) of the tripartite leader cDNA (Yueh and Schneider 1996). Inversions of the tripartite leader cDNA fragments are indicated according to the junction of the retained and inverted sequence. Inversions were generated by cloning the complimentary orientation by use of available restriction enzyme cleavage sites. Point mutations were introduced into the tripartite leader cDNA by overlap-extension PCR mutagenesis, by use of complementary oligonucleotides containing the altered sequence. All mutations were verified by DNA sequence analysis. Point mutations of the tripartite leader complementarity to 18S rRNA in each of the three elements (C1-C3) generated the following sequences relative to wild type (shown in Fig. 1): mutant C1 (nucleotide 5-AGAAUUCGU); mutant C2 (nucleotide 68-UCGCGUCUU); mutant C3 (nucleotide 157-CGUUUAGCUCU). Mutations were combined as indicated in the text. A stable hairpin was developed previously (Yueh and Schneider 1996) by linking five copies of a 12-nucleotide BamHI linker. The hairpin was inserted at position 220 in the tripartite leader cDNA. Full-length cDNA copies of the adenovirus IVa2 and IX 5' NCRs were prepared by reverse-transcription of viral mRNA with specific oligonucleotide primers. Human and Drosophila 5' NCRs were prepared by oligonucleotide-primed PCR from commercial HeLa cell and Drosophila cDNA libraries. Mutation of the IVa2 5' NCR (mIVa2) involved site-directed mutagenesis by PCR and excision of fragments to the boundaries shown in Figure 2. The IVa2 5' NCR deletion mutant (IVa2) was generated by deletion of sequences downstream of position 113. Mutation of the human hsp70 5' NCR involved deletion of the sequence downstream of position 96 (PstI; Huhsp70) or deletion of sequences between positions 150 and 215 (Hu mhsp70).

Transient transfections, and RNA and protein analyses

HEK 293 cells were grown in 100-mm dishes to ~50% confluence, transfected by the calcium phosphate procedure with 2 µg of reporter plasmid DNA and 0.5 µg of pGFP, which expresses a humanized green fluorescent protein to maintain transfection efficiencies. To determine the level of HBsAg synthesized and secreted, medium was removed at 40-48 hr following transfection and analyzed by commercial ELISA, as described by the manufacturer (Abbott Labs). The affect of adenovirus infection on HBsAg was determined by infection with 25 pfus per cell of adenovirus dl 309 18 hr post-transfection, medium was replaced at either 10 hr (prior to shutoff of cell protein synthesis) or 24 hr postinfection (after shutoff), followed by collection of medium 6 hr later and ELISA of HBsAg. The affect of heat shock on HBsAg synthesis was determined by replacement of medium at 36 hr post-transfection and heat shock of cells at 44°C for 4 hr, followed by collection of medium and ELISA of HBsAg. Data represent the average of five independent assays, with calculated standard errors shown. Northern blot RNA analysis was performed with total RNA purified from equal numbers of transfected cells. RNA was resolved by formaldehyde-agarose gel electrophoresis, transferred to blotting membrane and hybridized to ³²P-labeled DNA probes prepared against the HBsAg coding region.