rDNA-directed integration by an HIV-1 integrase--I-PpoI fusion protein

Abstract

Integrating viral vectors are efficient gene transfer tools, but their integration patterns have been associated with genotoxicity and oncogenicity. The recent development of highly specific designer nucleases has enabled target DNA modification and site-specific gene insertion at desired genomic loci. However, a lack of consensus exists regarding a perfect genomic safe harbour (GSH) that would allow transgenes to be stably and reliably expressed without adversely affecting endogenous gene structure and function. Ribosomal DNA (rDNA) has many advantages as a GSH, but efficient means to target integration to this locus are currently lacking. We tested whether lentivirus vector integration can be directed to rDNA by using fusion proteins consisting of the Human Immunodeficiency Virus 1 (HIV-1) integrase (IN) and the homing endonuclease I-PpoI, which has natural cleavage sites in the rDNA. A point mutation (N119A) was introduced into I-PpoI to abolish unwanted DNA cleavage by the endonuclease. The vector-incorporated IN-I-PpoIN119A fusion protein targeted integration into rDNA significantly more than unmodified lentivirus vectors, with an efficiency of 2.7%. Our findings show that IN-fusion proteins can be used to modify the integration pattern of lentivirus vectors, and to package site-specific DNA-recognizing proteins into vectors to obtain safer transgene integration.

Figure 1.

Production of lentivirus vectors and virus-like particles. (A) Lentivirus vector (LVV) and virus-like particle (VLP) production plasmids. The packaging plasmids were used either alone or mixed in equimolar amounts to generate LVVs and VLPs containing a single type of IN-molecule or mixed IN-multimers, respectively. (B) The IN-molecule content of LVVs was detected by an immunoblot using antiserum to HIV-1 IN. The vector-contained IN molecules on each lane are listed on the right. Expected molecular weights for IN/IN_D64V: 32 kDa; IN-I-PpoI_(N119A) 51 kDa. PRO, protease; RT, reverse transcriptase; RRE, Rev-responsive element; pA, polyadenylation signal; CMV, human cytomegalovirus immediate-early enhancer/promoter; cPPT, central polypurine tract; hPGK, human phosphoglycerate kinase promoter; GFP, green fluorescent protein; WPRE, Woodchuck hepatitis virus post-transcriptional regulatory element; SIN, self-inactivated LTR; LTR, long terminal repeat; RSV, Rous Sarcoma Virus promoter; VSV-G, Vesicular stomatitis virus G glycoprotein; IDLV, integration deficient lentivirus vector; Wt, wild type.

Figure 2.

The relative integration efficiency of different IN-modified vectors. HeLa cells transduced with LVVs containing different IN molecules were assayed by flow cytometry. For each vector the amount of GFP-positive cells in different time points was normalized to the value of day two, when GFP expression generally reaches its highest value (set to 100%). The integration efficiency of different vectors can be evaluated by looking at values after day 10, by which expression from unintegrated vector genomes has dropped close to zero.

Figure 3.

DSBs caused by LVV- or VLP- contained IN proteins. MRC-5 cells were transduced with LVVs (A) or VLPs (B) containing different IN molecules (middle). The DSB sites were detected by confocal microscopy using the antibody to γH2A.X (red). Nucleoli were labelled using the antibody to fibrillarin (green). Nuclei were visualized using DAPI (blue). The white arrowheads point to distinct shell-like nucleoli with adjacent DSB markers seen in cells transduced with the IN-I-PpoI-containing LVVs and VLPs. Hydrogen peroxide (H₂O₂; A) treatment was used as a positive control for DSBs. Scale bar: 5 µm. MRC-5, untreated MRC-5 cells (B). LVV, lentivirus vector; VLP, virus-like particle; DSB, DNA double-strand break.

Figure 4.

Cytotoxicity of IN-I-PpoI containing LVVs. HeLa (A) and MRC-5 (B) cells were transduced with two vector concentrations (2 and 10 ng of p24 per well) of LVVs containing different IN molecules (left). Cellular viability was measured 24, 48 and 72h after transduction (day 1, 2 and 3). Viability of the untreated cells at each time point is set to 100%. Viability of the vector-treated cells in a given time point is shown as the percentage of the untreated cells’ values. Differences between vector-treated groups and the untreated cell values were analysed at each time point using one-way ANOVA and the Dunnett’s Multiple Comparison Test. ***P < 0.001, **0.001 < P < 0.01, *0.01 < P < 0.05.

Figure 5.

Integration frequency in the rDNA. Integration frequency in the rDNA is shown. Previously published data sets of HIV-1 integration sites are shown as a reference. All vector sets were compared with each other. Differences between vectors were statistically significant only between IN_D64V+IN-I-PpoI_N119A and the rest of the vectors; the P-values for these comparisons are shown. IS, integration site. ^aHIV-1 vector integration sites (53) generated with the same restriction enzymes as the IN-fusion protein data sets. ^bHIV-1 vector integration sites generated with the restriction enzymes AvrII and MseI (43).

Figure 6.

Integration frequency in different genomic features. A heat map summarizes the relationships of vector integration site data sets (indicated above the columns) to selected genomic features (left of the corresponding row of the heat map). Tile colour indicates whether integration by different vectors is favored (increasing shades of red) or disfavored (increasing shades of blue) in a given feature relative to their matched random controls, as detailed in the colored receiver operating characteristic area scale at the bottom of the panel. The p-values shown as asterisks (*p < 0.05, **p < 0.01, ***p < 0.001) emerge from significant departures from the wt IN data set (53). The base pair values in the row labels indicate the size of the genomic interval used for analysis. Statistical methods and detailed naming of the genomic features: Berry et al. (30) and Brady et al. (33,54).