Fig 1.
Flowchart of library construction and screening.
The E. coli primary library was constructed by using a human genomic DNA fragment and the pΔBM-d2EGFP AscI vector plasmid, which contains IR/MAR, BSR, and d2EGFP expression cassettes. DNA from the library was transfected into CHO DG44 cells, and stable transformants showing high d2EGFP expression were isolated. The human genomic inserts were recovered from the cells and were used to construct the secondary E. coli library using the same IR/MAR vector (pΔBM-d2EGFP AscI). The library DNA was again transfected into CHO DG44 cells, and a secondary library showing high d2EGFP expression was identified. The plasmids were cloned from the secondary library and transfected into CHO DG44 cells. The cloned plasmid that showed higher d2EGFP expression than that of the vector plasmid was identified. The human genomic insert in this plasmid was sequenced. For details, see Methods.
Fig 2.
Effect of a palindrome sequence on gene expression from the amplified structure.
Effect of a 2 × 25 bp palindrome on gene expression from the repeated sequence. pΔBM-d2EGFP-AscI (containing the palindrome) or pΔBM-d2EGFP (no palindrome) was transfected into CHO DG44 cells. Transformants were selected with blasticidin for about 1 month, and then analyzed by flow cytometry. Results from two independent transfections are shown. Each panel shows the pΔBM-d2EGFP cell population (blue filled line) that overlapped with the pΔBM-d2EGFP-AscI cell population (unfilled line).
Fig 3.
Cloning of the B-3-31 sequence.
Plasmids from the positive secondary library were transfected into E. coli. Colony PCR followed by restriction enzyme digestion/gel electrophoresis revealed the composition of each secondary library (A). DNA from each clone was transfected into CHO DG44; transformants were selected for 1 month, and then analyzed by flow cytometry. In each panel, the test cell population (vector plus insert: blue filled line) overlaps with the control cell population (vector alone: unfilled line). Representative results from 11 of 21 clones examined are shown in B, and the remaining results appear in S1 Fig. Clone 31 from library B-3 (B-3-31) increased d2EGFP expression to the greatest extent, and this effect was reproduced in four independent transfections by three investigators (C).
Fig 4.
Effect of B-3-31 on IR/MAR-mediated gene amplification.
Metaphase chromosome spreads prepared from stable transformants were analyzed by FISH using a plasmid-derived probe. Representative images of small (S; panel A and C), medium (M; panel B), and large (L; panel C) HSRs are shown. The frequencies of cells bearing these HSRs among transformants from the control pΔBM-d2EGFP-AscI plasmid and the same plasmid containing the B-3-31 sequence were determined by examining more than 30 metaphase cells in each case from two independent transfections (panel D). If more than two HSRs were present in a cell, we counted it as a single cell bearing a HSR of the largest size.
Fig 5.
Effect of butyrate or 5-azacytidine on expression from the amplified plasmid array.
Thirty days after transfection with pΔBM-d2EGFP with or without B-3-31 into CHO DG44 cells, the cells were treated with the indicated concentration of butyrate or 5-azacytidine. After 3 days, d2EGFP expression was analyzed by flow cytometry. Charts on the left show the results obtained with pΔBM-d2EGFP cells, and charts on the right show the results obtained with cells harboring pΔBM-d2EGFP with B-3-31 (blue filled line) that overlapped with those harboring pΔBM-d2EGFP (unfilled line).
Fig 6.
Characteristics of B-3-31 sequence, and the dissection strategy.
(A) The panels show the locations of highly repetitive sequence (LINE and SINE), GC content, denaturation energy, locations of stress-induced duplex destabilization (SIDD) regions, predicted MAR potential, predicted nucleosome occupancy, inverted repeats, and 3D end-to-end distance of 480 bp sliding windows along the B-3-31 sequence. (B) To determine which region is responsible for elevation of gene expression, the indicated sub-regions were re-cloned and examined.
Fig 7.
The sub-regions of B-3-31 (bent, unbent, 3–1, 3–2, and 3–3; Fig 6B) were re-cloned into the AscI site of the IR/MAR vector (pΔBM-d2EGFP-AscI). The resultant plasmids and the control plasmid pΔBM-d2EGFP were transfected into CHO DG44 cells, and transformants were selected with BS. The results of flow-cytometric analysis of d2EGFP expression 30 and 51 days after transfection with the control pΔBM-d2EGFP are shown in A. In the chart, the cell population at 30 days (gray filled line) overlaps with the cell population at 51 days (unfilled line). Changes in average GFP intensity can be observed in the chart. At 51 days, the results of flow-cytometric analysis of d2EGFP expression in various cell populations are shown in B. In each chart, the test cell population (blue filled line) overlaps with the control cell population (pΔBM-d2EGFP; unfilled line). Average GFP intensities were observed for test cells, which can be compared with those of control cells at 51 days (panel A; 91.3). At days 30 and 51 after the transfection, average GFP intensities were divided by that of the control population, and the results are plotted in C. At day 29, cells were fixed and analyzed by FISH using a plasmid-derived probe. To analyze many cells, we evaluated gene amplification of the plasmid in interphase nuclei. Typical FISH images appear in D. Using these images as a standard, the frequency of each type of amplification was scored by examination of more than 300 nuclei in triplicate; the means +/- standard deviations are plotted in D.
Fig 8.
The B1, B2, and B3 sub-regions of B-3-31 (Fig 6B) were re-cloned into the AscI site of the IR/MAR vector (pΔBM-d2EGFP-AscI). The resultant plasmids and control plasmids with (pΔBM-d2EGFP) or without (pSFV-V d2EGFP) the IR/MAR were transfected into CHO DG44 cells, and transformants were selected with BS. The results of flow-cytometric analysis of d2EGFP expression at 36 and 49 days after transfection with the control pΔBM-d2EGFP plasmid are shown in A. In the chart, the cell population at 36 days (gray filled line) overlaps with the cell population at 49 days (unfilled line). At 49 days, the results of flow-cytometric analysis of d2EGFP expression in various cell populations are shown in B. In each chart, the test cell population (blue filled line) overlaps with the control cell population (pΔBM-d2EGFP; unfilled line). Average GFP intensities are noted for test cells, which can be compared to control cells at 49 days (panel A; 88.7). For two independent cultures from the same transfection, average GFP intensities obtained at the indicated times were divided by that of the control population (C). At day 31, the cells were fixed and analyzed by FISH using a plasmid-derived probe. The frequency of each type of amplification (representative image appears in Fig 7D) was scored by examining more than 300 nuclei in triplicate; the means +/- standard deviations are plotted in D.
Fig 9.
Frequency of the CpG dinucleotide along the amplified plasmid array.
The pΔBM-d2EGFP plasmid is amplified as a direct repeat, because it is an IR/MAR plasmid (dashed arrows above). The frequency of the CpG dinucleotide along the pΔBM-d2EGFP plasmid is plotted at the middle panel. The frequency of CpG along the B-3-31 sequence is shown in the lower panel, and it was inserted into pΔBM-d2EGFP plasmid at the position indicated in the figure.