Shortened nuclear matrix attachment regions are sufficient for replication and maintenance of episomes in mammalian cells

Abstract

Matrix attachment regions (MARs) can mediate the replication of vector episomes in mammalian cells; however, the molecular mode of action remains unclear. Here, we assessed the characteristics of MARs and the mechanism that mediates episomal vector replication in mammalian cells. Five shortened subfragments of β-interferon MAR fragments were cloned and transferred into CHO cells, and transgene expression levels, presence of the gene, and the episomal maintenance mechanism were determined. Three shortened MAR derivatives (position 781-1320, 1201-1740, and 1621-2201) retained full MAR activity and mediated episomal vector replication. Moreover, the three shortened MARs showed higher transgene expression levels, greater efficiency in colony formation, and more persistent transgene expression compared with those of the original pEPI-1 plasmid, and three functional truncated MARs can bind to SAF-A MAR-binding protein. These results suggest that shortened MARs are sufficient for replication and maintenance of episomes in CHO cells.

FIGURE 1:

Transfection efficiency (A) and transient expression (B) in transfected CHO cells. Plasmids with shortened MAR 1–5 were transfected into CHO cells, respectively. Transfection efficiency and eGFP transient expression were determined using flow cytometry after transfection for 48 h. Three independent experiments were performed in this study. SEM is indicated (Student’s t test,*P < 0.05). The results indicated that all shortened MAR could enhance the transfection efficiency and transient expression. The transfection efficiency of plasmid with the 540-base-pair DNA random sequence were no difference compared with shortened MAR.

FIGURE 2:

Establishment efficiency and eGFP expression in stably transfected CHO cells. (A) Establishment efficiency was determined by colony-forming assay. (B) The relation of transfection efficiency (%) and vector size (kb). There were significant differences among MAR-3,4,5, MAR-1,2, and full-length MAR. (C) Two weeks after transfection, CHO cells transfected with the shortened MAR 1–5 and full-length MAR were collected and the MFI of eGFP were examined by FACS Calibur. Mean values differed significantly between these plasmids (*P < 0.05).

FIGURE 3:

Long-term eGFP expression stability in transfected CHO cells. (A) The stably transfected CHO cells were cultured either with or without G418 selection pressure and MFI was detected to evaluate the stability of the expressed eGFP after 10, 18, 25, 33, 42, and 50 generations posttransfection, respectively. (B) The percentage of eGFP expressing cells in CHO cells transfected with full-length MAR, MAR-3, MAR-4, and MAR-5 vectors. (C) Retention of eGFP expression levels in cells transfected with full-length MAR, MAR-3, MAR-4, and MAR-5 vectors. (D) Fluorescence microscopy of eGFP gene in CHO cells transfected with full-length MAR, MAR-3, MAR-4, and MAR-5 grown with or without G418 selection pressure.

FIGURE 4:

Analysis of state of existence of the transgene in CHO cells and copy number. (A) Rescue experiments in E. coli with Hirt extract from CHO cells transfected with MAR-3, MAR-4, and MAR-5. M: DNA marker; lanes 1, 3, 5: original plasmids with MAR-3, MAR-4, and MAR-5 double-digested with KpnI and BamHI, respectively; lanes 2, 4, 6: rescued plasmids with MAR-3, MAR-4, and MAR-5 double-digested with KpnI and BamHI 50 generations after transfection, respectively. (B) Southern analysis of DNA isolated from CHO cells transfected with MAR-3, MAR-4, and MAR-5. The hybridization pattern of one representative clone is shown for each construct. M: DNA marker; lane 1, 3, 5: original plasmids with MAR-3, MAR-4, and MAR-5 digested with BamHI, respectively; lane 2, 4, 6: rescued plasmids with MAR-3, MAR-4, and MAR-5 digested with BamHI 50 generations after transfection, respectively. Lane 7: EcoRI digestion of untransfected CHO cells; lanes 8, 9: CHO cells transfected with pEGFP-C (chromosomal DNA was isolated, digested with EcoRI, separated on a 0.8% agarose gel, blotted, and hybridized with pEGFP-C1 probe). (C) FISH analysis of eGFP served as a probe in CHO cells transfected with vector no eGFP gene (negative control), no MAR (pEGFP-C1), full-length MAR, MAR-3, MAR-4, and MAR-5, respectively. The vector without eGFP gene can only see blue metaphase chromosomes, integration vector pECFP-C1 can see vector insert into chomosomes, vector with full-length MAR, MAR-3, MAR-4 and MAR-5 was episomal state on metaphase chromosomes (blue: metaphase chromosomes; red: vectors). (D) The gene copies of each metaphase plate as determined by FISH analysis. Fifty metaphase spreads were analyzed by FISH for each clone, an average vector copy number was estimated, and SEM is indicated. (E) The copy number was assessed by qPCR analysis. A serial dilution with a plasmid containing the eGFP gene was used to determine the absolute copy number. Three independent experiments were performed in this study. SEM is indicated (Student’s t test, *P < 0.05).

FIGURE 5:

ChIP and EMSA experiment. (A) PCR analysis of chromatin immunoprecipitation reactions, M: 500-base-pair marker; (B) ChIP-qPCR enrichment is shown as the percentage of input DNA; ChIP was performed on CHO/subfragment MAR-3,4, 5 using a specific antibody against SAF-A. Immunoprecipitated DNA was subjected to qPCR. (C) Representative of SAF-A binds specificity to subfragment MAR-4.

FIGURE 6:

Schematic representation of vectors, β-IFN MAR subfragment and illustration of its transcription factor binding motifs. (A) pEGFP-C1 (Clontech, Mountain View, CA) was used as the original vector. (B) β-IFN MAR (GenBank No: M83137.1) inserts into the downstream of eGFP reporter gene of the pEGFP-C1 plasmid to generate pEPI. (C) MAR-1, MAR-2, MAR-3, MAR-4, and MAR-5 represent the DNA fragments from the positions 1–480, 361–900, 781–1320, 1201–1740, and 1621–2201 of human β-IFN MAR; schematic illustration of various transcription factor binding motifs within the full-length human β-IFN MAR and five shortened β-IFN MAR.