The expression of genes delivered by retroviral vectors is often inefficient, a potential obstacle for their widespread use in human gene therapy. Here, we explored the possibility that the posttranscriptional regulatory element of woodchuck hepatitis virus (WPRE) might help resolve this problem. Insertion of the WPRE in the 3′ untranslated region of coding sequences carried by either oncoretroviral or lentiviral vectors substantially increased their levels of expression in a transgene-, promoter- and vector-independent manner. The WPRE thus increased either luciferase or green fluorescent protein production five- to eightfold, and effects of a comparable magnitude were observed with either the immediate-early cytomegalovirus or the herpesvirus thymidine kinase promoter and with both human immunodeficiency virus- and murine leukemia virus-based vectors. The WPRE exerted this influence only when placed in the sense orientation, consistent with its predicted posttranscriptional mechanism of action. These results demonstrate that the WPRE significantly improves the performance of retroviral vectors and emphasize that posttranscriptional regulation of gene expression should be taken into account in the design of gene delivery systems.

The hepatitis B virus posttranscriptional regulatory element (HBVPRE) is a cis-acting RNA element that partially overlaps with enhancer I and is required for the cytoplasmic accumulation of HBV surface RNAs. We find that the closely related woodchuck hepatitis virus (WHV), which has been shown to lack a functional enhancer I, also contains a posttranscriptional regulatory element (WPRE). Deletion analysis suggests that the WPRE consists of three independent subelements. Comparison of the bipartite HBVPRE and tripartite WPRE activities reveals that the tripartite WPRE is two to three times more active than the bipartite HBVPRE. Mutation of a single WPRE subelement decreases WPRE activity to the level of the HBVPRE. Bipartite and tripartite chimeras of the WPRE and HBVPRE possess activities which suggest that elements containing three subelements are posttranscriptionally stronger than those containing two. These data demonstrate that the posttranscriptional regulatory element is conserved within the mammalian hepadnaviruses and that its strength is determined by the number of subelements within the RNA.

1 The polymerase chain reaction (PCR) was used in combination with plaque hybridization analysis to clone four variants of the EP3 prostaglandin receptor from a human small intestine cDNA library. 2 Three of these variants, i.e. the EP3A, EP3E and EP3D, share the same primary amino acid sequence except for their carboxyl termini, which diverge from one another at the same point, approximately 10 amino acids away from the end of the seventh membrane spanning domain of the receptor. The fourth variant (EP3A1) has a nucleotide coding sequence identical to EP3A but has a completely different 3' untranslated sequence. 3 The carboxyl termini of the three isoforms differ most obviously in length with the EP3A being the longest (41 amino acids) and the EP3E being the shortest (16 amino acids). They also differ in content with the EP3A containing 9 serine and threonines in its carboxyl terminus and the EP3E none. 4 Transient expression in eukaryotic cells showed that the human EP3 receptor variants had similar but not identical radioligand binding properties and differed in their functional coupling to second messenger pathways. Up to 3 pmol mg-' protein of [3H]-prostaglandin E2 binding could be obtained with more than 95% specific binding. Using a reporter gene assay, as a measure of intracellular cyclic AMP levels, the EP3A coupled more efficiently to the inhibition of adenylyl cyclase than did the EP3E. 5 PCR was used to confirm the presence of mRNAs encoding the four human EP3 receptor variants in tissues of the human small intestine, heart and pancreas. These findings indicate that the EP3 receptor variants identified here are likely to be expressed in tissues. The differences in the carboxyl termini at the protein level, and in the 3' untranslated regions at the mRNA level, could be profound in terms of the regulation and functional coupling of these receptor isoforms.

Viruses often contain cis-acting RNA elements, which facilitate the posttranscriptional processing and export of their messages. These elements fall into two classes distinguished by the presence of either viral or cellular RNA binding proteins. To date, studies have indicated that the viral proteins utilize the CRM1-dependent export pathway, while the cellular factors generally function in a CRM1-independent manner. The cis-acting element found in the woodchuck hepatitis virus (WHV) (the WHV posttranscriptional regulatory element [WPRE]) has the ability to posttranscriptionally stimulate transgene expression and requires no viral proteins to function. Conventional wisdom suggests that the WPRE would function in a CRM1-independent manner. However, our studies on this element reveal that its efficient function is sensitive to the overexpression of the C terminus of CAN/Nup214 and treatment with the antimicrobial agent leptomycin B. Furthermore, the overexpression of CRM1 stimulates WPRE activity. These results suggest a direct role for CRM1 in the export function of the WPRE. This observation suggests that the WPRE is directing messages into a CRM1-dependent mRNA export pathway in somatic mammalian cells.The generation of mature cytoplasmic mRNAs requires numerous processing steps, namely, transcription, capping, splicing, polyadenylation, and transport to the cytoplasm. This process is tightly regulated, since aberrant transcripts are degraded in the nucleus and only properly processed mRNAs are exported to cytoplasm (43). Thus, nuclear export ensures that only completely processed mRNAs can be translated into protein.Much of our present understanding of nuclear export has come from the study of how viruses exploit host cell RNA processing and export pathways. The first viral export system studies were those of complex retroviruses, exemplified by human immunodeficiency virus type 1 (HIV-1). HIV-1 replication requires unspliced and partially spliced RNAs to be exported from the nucleus by the virally encoded Rev protein (9,13,44). Rev contains an RNA binding domain, which specifically binds to the Rev response element (RRE), located within the second intron of HIV-1 pre-mRNA, and a nuclear export signal (NES) that interacts with CRM1, a member of the importin ␤ family of transport receptors (6,15,17,44,58).The interaction between Rev and CRM1 is dependent upon CRM1 association with the GTP-bound form of the GTPase Ran protein (RanGTP). Once assembled, the RRE/Rev-CRM1-RanGTP ribonucleoprotein complex interacts with NPs, which trigger its nuclear export. CRM1 has been proposed to mediate this interaction by directly contacting selected nucleoporins (NPs), including CAN/Nup214. Binding of CRM1 to CAN has been mapped to the NP domain located within the extreme carboxy terminus of CAN (16). Overexpression of the isolated NP domain of CAN, termed ⌬CAN, is able to inhibit Rev-mediated export by competing with the NPs for binding to CRM1 (2). Rev-mediated export is also inhibited by the antibiotic leptomycin B (LMB), w...