Integration by Design

Integration by design Suzanne Sandmeyer* Department of Biological Chemistry, College of Medicine, University of California, Irvine, CA 92697-1700 s goes history, so goes research: genomes are broadly accessed by retro- ends of the full-length DNA mediates this year, activity in areas of viruses, but that there are decidedly integration into host DNA. Isolation retrovirus research related only nonrandom patterns as well (8). More first of preintegration complexes from Aindirectly have provoked events recently, large numbers of HIV type 1 infected cells and then production of that are notable when considered to- (HIV-1) insertions have been mapped and active, recombinant IN allowed exami- gether. Last summer it was reported compared with genomewide transcrip- nation of the effect of different target that a patient in one X-linked severe tion patterns to globally probe the rela- features on integration in vitro. A gener- combined immunodeficiency retroviral tionship between gene expression and alization that has emerged from studies vector gene therapy trial had developed retrovirus integration (9). These experi- conducted in several laboratories is that leukemia. Now disquietingly, there has ments showed that HIV-1 insertion fa- bending of DNA favors integration (18), been a second such event, and a third vors transcribed regions. Nonetheless, as do hairpin structures (19). The patient is reported to have a vector in- the basis of the preference for tran- former occurs in nucleosomes, which, sertion near the same gene (LMO2)as scribed regions has been elusive, and contrary to expectations, were found to observed in the other two individuals examination of at least one transcribed act as preferred targets over nonnucleo- (1). Meanwhile, in a basic research labo- region for effects of transcriptional ac- somal DNA, both in vitro and in vivo ratory, experiments have moved us an- tivity on integration activity have not (20–22). other step closer to understanding the shown a positive correlation (10). The relatively global distribution of mechanics of insertion specificity for At the heart of retroviral integration retrovirus integration sites stands in retrovirus-type integrases (IN). As re- is the IN. It is a member of the interesting contrast to the distinctive ported in this issue of PNAS, investiga- D,D(35)E transposase͞IN superfamily insertion preferences of their LTR- tors have produced active retroviruslike named after its conserved catalytic triad retrotransposon cousins, the Pseudo- elements with synthetic insertion speci- viridiae (e.g., Ty1 and Ty5 copialike ele- ficities (2). Dan Voytas and colleagues ments) (23) and the Metaviridae (e.g., at Iowa State University (Ames) study Potential deleterious Tf1 and Ty3 gypsylike elements) (24). the Saccharomyces long terminal repeat IN proteins encoded by these elements (LTR)-retrotransposon Ty5, which tar- retrovirus insertions have the zinc-binding motif, the highly gets heterochromatic regions (3). Now, conserved residues of the central do- in an elegant adaptation of the two-hy- fueled investigation main and the poorly conserved C-termi- brid system, the 6-aa Ty5 targeting do- nal domain. The IN proteins of the main (TD) was exchanged for two heter- into the mechanism of Pseudoviridae and the Metaviridae differ ologous domains shown to mediate from each other in the C-terminal do- interaction of their respective proteins insertion site selection. main where the Pseudoviridae have a conserved GKGY motif (23), and the with protein partners. When domains ͞ from those partners were produced Metaviridae have a conserved GPF Y motif. Some members of the Metaviridae fused to the LexA DNA-binding do- of amino acids. Because of its central also have a chromodomain (24). main, targeting to LexA-binding sites role in the retrovirus lifecycle, the func- As a group, the yeast LTR retrotrans- was observed. Although integration tion and structure of this enzyme has posons have notable insertion prefer- specificity in the system was by no been studied extensively (reviewed in ences. The specificity of Ty5 for hetero- means absolute, these results are of in- refs. 11–15). Retroviral IN mediates a chromatin is discussed further below. In terest to genetic engineers and future strand transfer of LTR DNA 3Ј OH budding yeast, the Pseudoviridae Ty1, 2, gene therapists. ends to staggered positions in the host and 4 reside mostly within 750 bp of the DNA (16, 17). Combined evidence of Interest in the integration patterns of 5Ј ends of tRNA genes (25, 26). In vivo retroviruses is longstanding. Despite the many types shows a retroviral IN with insertions fall along a gradient begin- potential danger of deleterious activat- three physically distinct domains. An Ϫ Ј ␣ ning at about 80 bp from the 5 cod- ing or even inactivating insertions, retro- N-terminal domain includes three - ing end of the tRNA gene and extend- viruses present compelling advantages as helices and a zinc-binding motif. This ing upstream. Integration appears to rise therapy vectors (reviewed in ref. 4). domain has been implicated in dimeriza- and fall in a pattern which could corre- Early investigations of oncogenic retro- tion and in binding the LTR ends. The late with some feature of the nucleo- virus insertion sites in transformed cells central domain contains the conserved some (27). The pattern of integration of showed that insertions were linked to catalytic triad D,D(35)E. Members of the Metaviridae element Ty3 is even activation of flanking oncogenes or this triad coordinate a divalent metal more restricted. The gene-proximal 2ϩ DNaseI hypersensitive sites, leading to cation, probably Mg in vivo (15) and strand transfer in this case occurs within the notion that insertion into open chro- are essential for catalytic activity. The one or two nucleotides of tRNA gene matin was favored (reviewed in ref. 5; C-terminal domain contributes to oli- transcription initiation sites. In vivo it is see also refs. 6 and 7). The potential for gomerization, has nonspecific DNA- likely that transcription factors TFIIIB deleterious retrovirus vector insertions binding activity and is physically similar and TFIIIC are essential for Ty3 target- fueled investigation into the mechanistic to the SH3 protein interaction domain. ing (28–30). Furthermore, it has been basis of insertion site selection. Devel- No full-length IN structure has yet been opment of PCR assays with which signif- determined at high resolution. icant numbers of retrovirus integration In vivo a retroviral preintegration See companion article on page 5891. sites could be mapped showed that complex composed of IN bound to the *E-mail: [email protected]. 5586–5588 ͉ PNAS ͉ May 13, 2003 ͉ vol. 100 ͉ no. 10 www.pnas.org͞cgi͞doi͞10.1073͞pnas.1031802100 Downloaded by guest on September 29, 2021 COMMENTARY shown that yeast elements Ty1–4 target other genes transcribed by RNA poly- merase III with similar patterns to those observed flanking tRNA genes (27, 30). In vitro, Ty3 targeting to the U6 gene requires only TATA-binding protein and Brf1 (29). Observation of highly specific integration in yeast helped to motivate a series of experiments to confer novel insertion specificities on retrovirus IN proteins (reviewed in refs. 31 and 32). Recombi- nant retroviral IN has been expressed as a fusion with relatively compact DNA- binding domains including lambda re- pressor (33), LexA DNA-binding domain (34, 35), and the DNA-binding domain of Zif268 (36). Recombinant proteins have been shown to target in vitro integration to the respective DNA- binding sites of the fusion proteins. Dis- appointingly, these chimeric IN species, appear to be incompatible with high lev- els of infectious virus. Presumably this is caused by some failure to structurally Fig. 1. Strategy for retargeting Ty5 integration. Top, schematic of Ty5 single ORF encoding RNA binding accommodate the heterologous domain. (RB), protease (PR), integrase (IN), reverse transcriptase (RT), and marker gene (his3AI) (open box). View To circumvent some of these problems, of IN is expanded to show conserved residues and targeting domain (TD) (solid). Lower left, preintegration a strategy involving trans expression of complex showing wild-type IN bound to ends of Ty5 DNA (thick line) and integrating into telomeric IN has been used. In this variation, a heterochromatin, mediated by Sir4p (hatched). Lower right, same as left except that the natural TD is fusion of HIV-1 structural protein p6 to replaced with heterologous domains (TD*) from Rad9p and NpwBP (solid). LexA DNA-binding domain an IN-LexA targeting domain directs IN (open) is expressed fused to Sir4pC, Rad53p FHA1, and Npw38 WW domains (hatched). Integration occurs proximal to LexA-binding sites (open arrows) in plasmid target (closed circle). to the virion and complements catalyti- cally defective IN contributed from Gag- Pol (37, 38). However, there are no nat- ruption of IN. A 13-aa sequence in tion of the majority of (nontarget plas- urally occurring LexA-binding sites in Rad9p mediates its interaction with a mid) Ty5 integrations? Do nonplasmid mammalian cells, and targeting to syn- forkhead-associated domain (FHA1) in insertions default to random, to native thetic sites has not yet been reported. another DNA repair protein, Rad53p. A Rad53p direction in the case of the Ty5 is distinct among the yeast ele- 12-aa domain in NpwBP mediates inter- Rad9p-based TD, or do natural, as yet ments. Originally identified as a degen- action with the WW domain of another unidentified, functions continue to oper- erate element at the ends of Saccharo- nuclear protein Npw38. The Rad9p and ate on the Ty5 IN? Is it possible to gen- myces cerevisiae chromosomes (39), the NpwBP domains were substituted for erate integration that is more highly Voytas laboratory recovered an active the natural Ty5 TD. The partner inter- restricted, perhaps through the use of copy from Saccharomyces paradoxus and acting domains (i.e., FHA1 from phage panning or slightly larger transferred it into S.

Load more