Regulation And Function Of Retroelements

Child Health And Human Development

Investigators

Linked publications & trials

Abstract

Diseases caused by retroviruses such as HIV-1 and the adaptation of retroviruses for use in gene therapy has intensified the need to understand how these viruses replicate. Primary objectives of ours are to understand how reverse transcription of viral mRNA occurs and how the cDNA products are integrated into the genome of infected cells. As a result of their similarity to retroviruses, LTR-retrotransposons are important models for retrovirus replication. The retrotransposon we study is the Tf1 element of the fission yeast, Schizosaccharomyces pombe. Tf1 efficiently reverse transcribes its mRNA and its integrase (IN) inserts the cDNA into the genome of S. pombe. Our studies of reverse transcription and integration are greatly facilitated by the techniques of yeast genetics. This approach not only allows us to identify features of the transposon critical for activity but host genes required for transposition can also be identified. The ease of culturing yeast also allows us to apply biochemistry to investigate the mechanisms we identify by genetics. The integration of HIV-1 cDNA shows a significant preference for actively transcribed genes. Similarly, the insertion of murine leukemia virus shows a strong preference for sites within 5kb of transcription initiation. Very little is known about how these viruses interact with the structures of chromatin and recognize transcription units. Our recent observation that the integration of Tf1 occurs specifically at pol II promoters presents the real opportunity to study in S. pombe an integration mechanism that parallels that of retroviruses. This year we explored the ability of Tf1 to recognize pol II promoters by creating an integration assay for specific target plasmids. We generated a plasmid with the ade6 gene of S. pombe and included it in a strain that was induced for Tf1 transposition. Fifty separate insertions into the target plasmid were isolated and analyzed. Ninety-five percent of the inserts occurred within a 160 nucleotide window in the promoter of ade6. This target assay clearly reproduced the promoter preference we observed with insertions into genomic sites. The insertions within the 160 nucleotide window exhibited a strong pattern of periodicity. Four narrow clusters of inserts were spaced approximately 30 bp apart. The window of 160 nucleotides corresponded with the amount of DNA wrapped around a single nucleosome and consistent with the possibility, the 30 bp pattern corresponded to the amount of DNA bound to the histone fold pairs within a single nucleosome. The IN of Tf1 contains Zn finger and catalytic domains similar to INs of retroviruses. Unlike other INs, the Tf1 IN possesses a chromodomain at its C-terminus. Chromodomains have been found to bind directly to histone H3 in specific nucleosomes. This suggests the possibility that Tf1 integration is mediated by an interaction between IN and specific nucleosomes at pol II promoters. Consistent with this model, we found that IN purified from bacteria bound specifically to histone H3 and not H2a, H2b, or H4. In addition, mutations in the chromodomain caused a significant reduction in transposition and a defect in recognition of the ade6 promoter. Other evidence that the chromodomain recognizes histones is that the insertions we isolated in the promoter of ade6 corresponded to a location where published data indicates a positioned nucleosome exists. To study the function of the chromodomain in vitro we developed biochemical assays for IN activity. We attached a six-his tag to the N-terminus of Tf1 IN and purified the protein from E. coli using Ni resin. The INs of retroviruses including that of HIV-1 are extremely insoluble and difficult to concentrate. We were surprised to find that the full-length IN of Tf1 was easy to purify and could readily be isolated in high concentrations. The INs of retroviruses possess processing activity that removes two terminal nucleotides from the 3? ends of the cDNA before insertion occurs. Based on the position of the minus strand primer, LTR-retrotransposons are not believed to possess processing activity. However our sequence data of Tf1 cDNA revealed 95% of the 3? ends had untemplated addtions. A 3? processing activity would remove these nucleotides and position the critical ?CA? dinucleotide at the ends of the cDNA where they must be located for integration to occur. We assayed the Tf1 IN for processing activity using oligonucleotides that mimic the ends of the cDNA. The IN had strong processing activity that removed two, three, four or five additional nucleotides from the 3? end of the oligos. This processing activity as observed in vitro suggests that the multiple nucleotides present the both ends of the cDNA could be removed by IN in vivo. This would allow integration of the most prevalent cDNA species we detected. We also tested the Tf1 IN for integration activity. We found the IN did have strong IN activity as indicated by the ability of the enzyme to insert oligonucleotides into each other. The resulting products were substantially larger than the original oligonucleotides. Oligonucleotides that mimicked the donor cDNA were altered to test whether Tf1 IN required specific sequences at the 3? ends. As is the case with other INs, Tf1 IN had a strong requirement for the dinucleotide CA at the 3? ends. The results of the processing and integration assays demonstrated that the IN of Tf1 has the same activities as retroviral INs and therefore is an excellent model for the IN of HIV-1. In addition, the high solubility of the Tf1 IN suggests that the protein may form crystals that could result in the first high resolution structure of an intact IN. This structure could contribute significantly to our understanding of retrovirus integration. We are pursuing this possibility in collaboration with the lab of David Davies. To explore the function of the chromodomain we expressed in E. coli Tf1 IN that lacked the chromodomain. This chromo minus IN (CH-) was assayed for integrase activities using the same artificial substrates described above. CH- possessed strong activities in both the processing and integration reactions. We were surprised to find that the enzyme lacking the chromodomain was about 10-fold more active that the intact IN. We also tested whether the chromodomain contributed to the recognition of the terminal dinucleotide at the 3? end of the donor DNA. The CH- protein exhibited a surprising relaxation of the sequence requirement at the ends of the donor DNA. Taken together, these results indicated that the chromodomain functioned as a negative regulator of integration and as a specificity factor for the donor DNA. One possibility that we are currently testing is that the chromodomain inhibits integration until it identifies a specific nucleosome at a pol II promoter. The chromodomain then binds H3 of that nucleosome. This alters the conformation of IN in a way that stimulates integration activity.