Plasmid DNA sequences present in conventional herpes simplex virus amplicon vectors cause rapid transgene silencing by forming inactive chromatin

Abstract

The herpes simplex virus (HSV)-based amplicon vector, a bacterial-viral-mammalian cell shuttle system, holds promise as a versatile gene delivery vehicle because of its large transgene capacity. However, amplicon-mediated transgene expression is often transient. We hypothesized that the presence of prokaryotic DNA sequences within the packaged vector genome can trigger transcriptional silencing of the entire vector sequence. To test this, we constructed a novel amplicon vector devoid of bacterial sequences (minicircle [MC] amplicon). Although the same dose of the minicircle amplicon vector in normal human fibroblasts resulted in an expression of luciferase approximately 20 times higher than that caused by the conventional amplicon vector, no significant difference was observed in copy numbers of luciferase DNA between MC amplicon- and control-transduced cells. Quantitative analyses of levels of luciferase mRNA revealed that differential expression of luciferase was controlled at the transcriptional level. Chromatin immunoprecipitation PCR analyses of several regions of vector genomes revealed that the bacterial sequences found in the conventional amplicon DNA were associated with an inactive form of chromatin immediately after infection. The presence of bacterial sequences also affected the remaining vector sequences in the conventional amplicon vector. Finally, nude mice injected with the MC amplicon exhibited higher and more sustained expression of luciferase than those injected with the conventional amplicon, confirming the usefulness of the MC amplicon devoid of bacterial sequences. Although additional improvements are absolutely required, these findings are a significant first step toward developing a novel HSV amplicon vector that can achieve enhanced long-term transgene expression.

FIG. 1.

Production of the MC amplicon vector using φC31 integrase-mediated recombination in Escherichia coli. (A) Flow chart of φC31 integrase-mediated intramolecular recombination of pBAD.φC31.HGCag-Luc. The resulting products are shown and designated as MC-HGCag-Luc and φC31-backbone. MC-HGCag-Luc DNA was purified from contaminated pBAD.φC31.HGCag-Luc and φC31-backbone by digesting with KpnI and NotI followed by CsCl gradient ultracentrifugation. (B) Electrophoretic analysis of the MC amplicon, its precursor, and intermediates. One hundred nanograms of each DNA sample was digested with BglII and NotI and electrophoresed on a 0.8% agarose gel. Lane 1, pBAD.φC31.HGCag-Luc before induction; lane 2, DNA sample obtained after a 2-hour induction of φC31 integrase expression (mixture of pBAD.φC31.HGCag-Luc, MC-HGCag-Luc, and φC31-backbone); lane 3, purified MC-HGCag-Luc; and lane M, 1-kb-plus DNA ladder (Invitrogen). (C) Map of pHGCag-Luc amplicon vector plasmid used for a control in the study.

FIG. 2.

Time course of Luc activity in MRC9 human fibroblasts infected with either MC (⧫) or conventional HGCag (▪) amplicon vectors in culture. Luc activity is presented as mean ± standard deviation (SD) (n = 4).

FIG. 3.

Time course of vector genome copy numbers in MRC9 human fibroblasts infected with either MC (⧫) or conventional HGCag (▪) amplicon vectors in culture. luc gene copy numbers were determined by real-time PCR analysis. Data are presented as mean ± SD (n = 4).

FIG. 4.

Time course expression of luc mRNA in MRC9 human fibroblasts infected with either MC (⧫) or conventional HGCag (▪) amplicon vectors in culture. Each value was calculated relative to that of the MC amplicon at day 1. Data are presented as mean ± SD (n = 3).

FIG. 5.

Time course modifications in chromatin of various regions of vector genomes in MRC9 human fibroblasts infected with either MC or conventional HGCag amplicon vectors. (A) Schematic structures of MC and HGCag amplicon vector genomes and approximate locations of the seven primer pairs examined. (B) Ratio of DNA associated with Met-K9-H3 to that with Met-K4-H3 was calculated for each of the seven genetic regions shown in panel A. The higher the ratio, the more inactive the transcription in the region.

FIG. 6.

Time course of Luc expression in nude mice in vivo. (A) Nude mice were injected systemically via the tail vein with either MC or HGCag amplicon vectors, and Luc expression was monitored with a bioluminescent imaging system. Numbers shown beneath photos are photon counts. (B) Total photon counts were calculated from the acquired images using CMIR image software, and data are presented as means ± SD (n = 6). The data for animals injected with the MC and HGCag amplicons are plotted as ⧫ and ▪, respectively. RLU, relative light units. The horizontal broken line indicates the background level of the bioluminescent signal.

FIG. 6.

Time course of Luc expression in nude mice in vivo. (A) Nude mice were injected systemically via the tail vein with either MC or HGCag amplicon vectors, and Luc expression was monitored with a bioluminescent imaging system. Numbers shown beneath photos are photon counts. (B) Total photon counts were calculated from the acquired images using CMIR image software, and data are presented as means ± SD (n = 6). The data for animals injected with the MC and HGCag amplicons are plotted as ⧫ and ▪, respectively. RLU, relative light units. The horizontal broken line indicates the background level of the bioluminescent signal.