Efficient assembly of de novo human artificial chromosomes from large genomic loci

Abstract

Background: Human artificial chromosomes (HACs) are potentially useful vectors for gene transfer studies and for functional annotation of the genome because of their suitability for cloning, manipulating and transferring large segments of the genome. However, development of HACs for the transfer of large genomic loci into mammalian cells has been limited by difficulties in manipulating high-molecular weight DNA, as well as by the low overall frequencies of de novo HAC formation. Indeed, to date, only a small number of large (>100 kb) genomic loci have been reported to be successfully packaged into de novo HACs.

Results: We have developed novel methodologies to enable efficient assembly of HAC vectors containing any genomic locus of interest. We report here the creation of a novel, bimolecular system based on bacterial artificial chromosomes (BACs) for the construction of HACs incorporating any defined genomic region. We have utilized this vector system to rapidly design, construct and validate multiple de novo HACs containing large (100-200 kb) genomic loci including therapeutically significant genes for human growth hormone (HGH), polycystic kidney disease (PKD1) and beta-globin. We report significant differences in the ability of different genomic loci to support de novo HAC formation, suggesting possible effects of cis-acting genomic elements. Finally, as a proof of principle, we have observed sustained beta-globin gene expression from HACs incorporating the entire 200 kb beta-globin genomic locus for over 90 days in the absence of selection.

Conclusion: Taken together, these results are significant for the development of HAC vector technology, as they enable high-throughput assembly and functional validation of HACs containing any large genomic locus. We have evaluated the impact of different genomic loci on the frequency of HAC formation and identified segments of genomic DNA that appear to facilitate de novo HAC formation. These genomic loci may be useful for identifying discrete functional elements that may be incorporated into future generations of HAC vectors.

Figure 1

Strategy for construction of a bimolecular, prefabricated, linear HAC vector. Digestion of BAC-CEN and BAC-GEN vectors with the ultra-rare homing endonucleases I-CeuI and PI-SceI permits directional ligation of both "arms" to form a linear HAC vector.

Figure 2

Pulsed Field Gel Electrophoresis showing creation of HAC vector. Ligation reactions were set up using linearized BAC-CEN alone (Lane 1), linearized BAC-GEN alone (Lane 2), or linearized BAC-CEN and BAC-GEN together (Lane 3). A clear ligation product (arrow), the linear HAC vector, is only visible in Lane 3. Note that trace amounts of re-circularized CEN and GEN arms are detectable in Lanes 1 and 2, but these do not significantly affect assembly or purification of the HAC.

Figure 3

Cytogenetic validation of de novo HAC vectors containing the ß-globin genomic locus. A) D17Z1 alpha-satellite (green), BAC vector (red). (B) D17Z1 alpha-satellite (green), ß-globin genomic locus (red). (C) D17Z1 alpha-satellite (green), telomere DNA (red). (D) D17Z1 alpha-satellite (green), CENP-C (red). In all cases, DNA is in blue (DAPI). Arrows point to de novo HACs.

Figure 4

Analysis of ß-globin mRNA expression from cell lines containing de novo ß-globin HACs after 30 days in the absence of selection. Poly A+ RNA from individual cell clones was subjected to first strand synthesis and PCR with primers specific to exon-3 of the ß-globin gene. Arrow indicates the ß-globin PCR product. 1) Untransfected HT1080. (2) Untransfected HT1080, no reverse transcriptase (RT). (3) ß-globin HAC clone #1, +RT. (4) ß-globin HAC clone #1, -RT. (5) ß-globin HAC clone #2, +RT. (6) ß-globin HAC clone #1, -RT. (7) ß-globin genomic DNA control