Systematic targeted integration to study Albumin gene control elements

Abstract

To study transcriptional regulation by distant enhancers, we devised a system of easily modified reporter plasmids for integration into single-copy targeting cassettes in clones of HuH7, a human hepatocellular carcinoma. The plasmid constructs tested transcriptional function of a 35-kb region that contained the rat albumin gene and its upstream flanking region. Expression of integrants was analyzed in two orientations, and compared to transient expression of non-integrated plasmids. Enhancers were studied in their natural positions relative to the promoter and localized by deletion. All constructs were also analyzed by transient transfection assays. In addition to the known albumin gene enhancer (E1 at -10 kb), we demonstrated two new enhancers, E2 at -13, and E4 at +1.2 kb. All three enhancers functioned in both transient assays and integrated constructs. However, chromosomal integration demonstrated several differences from transient expression. For example, analysis of E2 showed that enhancer function within the chromosome required a larger gene region than in transient assays. Another conserved region, E3 at -0.7 kb, functioned as an enhancer in transient assays but inhibited the function of E1 and E2 when chromosomally integrated. The enhancers did not show additive or synergistic behavior,an effect consistent with competition for the promoter or inhibitory interactions among enhancers. Growth arrest by serum starvation strongly stimulated the function of some integrated enhancers, consistent with the expected disruption of enhancer-promoter looping during the cell cycle.

Figure 1. Cloning and recombination strategies.

A: Rat Alb and its upstream region. Using restriction sites in the pLL1 linker, GFP was combined with Alb123, Region 1 (a NarI – XhoI segment from −13.7 to −3.9 kb), Region 2 (a segment from −3.9 to −0.2 kb, generated by PCR that added terminal FseI and PstI sites), and Alb123 (−0.2 to the transcription start, with terminal PstI and BglII sites) in a series of reporter plasmids. The 16.5-kb Alb gene was cloned by joining a 4.4 kb proximal segment (transcription start to SmaI, amplified by PCR that added an SgfI site) and a 12.1 kb distal segment (SmaI – EagI). To insert the Alb123-GFP reporter as an FseI–SmaI segment, the Alb-containing plasmid was cut with FseI and SgfI, and blunted at the latter site. The resulting reporter plasmids ranged from 4–40 kb (Table S1). B: Schematic map of the 290 bp linker showing the restriction sites for locus assembly. C: Mechanism of RMCE. The targeting cassette encodes an HY-TK fusion protein that makes the cell sensitive to Gancyclovir. Integration of reporter constructs in pLL1 or pLL2 occurs via Cre-mediated recombination, and can take place in two orientations due to the inverted arrangement of the LoxP sites . Loss of the HY-TK gene renders the integrant cells resistant to Gancyclovir.

Figure 2. Characterization of integrated constructs.

A: Chromosomal location of the targeting cassette in HuH7-9. Maps show representative integration of the 1.2-kb plasmid P. Plasmids may integrate in two orientations (O1, O2) between RELN and ORC5, respectively, in the same or opposite direction as transcription of both genes (arrows). PCR using primers R1 and F2 (from LoxP), and F1 and R2 (from flanking chromosomal regions), demonstrated that the integrant ends were intact and determined orientation. B: Representative PCR screening of integrants. O1 amplimersare F1 – R1 (302 bp) and F2 – R2 (599 bp). O2 amplimersare F1 – R2 (347 bp) and F2 – R1 (554 bp). C: Representative maps showing predicted Southern blot bands of integrated constructs P and 1+2+P obtained with KpnI (K) and HpaI (H). The hybridization probe is marked as a bar over the promoter-GFP region, and the promoter and known enhancer are marked in black. D: Southern blot mapping of eight integrated constructs. The blots (left,KpnI; right, HpaI digests) were probed with a 1.2 kb FseI - SmaI segment of plasmid P. The expected band sizes are listed in Table S3.

Figure 3. Transient transfection vs. stable gene integration for determining enhancer activity.

A: Plasmid constructs. Plasmid P contains the 0.2 kb promoter region fused to GFP. Plasmid 1+2+P also contains a 17.4 kb contiguous upstream segment (Regions 1 and 2). The dotted line represents the vector backbone. B: Transient transfection assays. Upper panel,scatter plots; lower panel, histogram subtraction. The solid green curve is the calculated fluorescence due to transfection. C: Plot of mean fluorescence intensity for each plasmid corrected for molar plasmid concentration and normalized to the promoter value. Average and standard deviation of 3 separate transfection experiments, each with triplicate transfections. D: Cell images. The promoter region produced weak gene expression that was strongly stimulated by the added enhancer region. The images show Orientation 1 integrants. E: Flow cytometric analysis of stable clones showing orientation-dependent stimulation by enhancers. Control cells, open curve; P, gray; 1+2+P, black. F: Expression in HuH7-9 cells. G: Expression in HuH7-10 cells. F, G; Plots show the average and standard deviation of fluorescence from three clones in each orientation. O1, O2; orientations 1 and 2.

Figure 4. Survey of a 35-kb region for novel regulatory elements.

A: Map of plasmid constructs. Conserved noncoding regions, black; Alb gene exons, gray; E1, the previously characterized Alb enhancer. B: Transient transfection analysis. Mean fluorescence intensity and standard deviation werecalculatedfrom triplicatetransfections in at least three separate experiments. C: Expression of stable integrants in O1. D: Expression of stable integrants in O2. C, D: Mean and standard deviation of expression from at least 2 integrant clones per construct in each orientation, except for a single integrant of 1+2+P+ALB in O1. Fluorescence intensity was normalized to the value obtained for the promoter.

Figure 5. Localization of E2, a novel upstream Alb enhancer.

A: Maps of deletion constructs. Large deletions of 1+2+P (top) were constructed with restriction enzymes ApaI and BstXI. Plasmid E2+P was further deleted by PCR (bottom; see Table S2). Black, regions highly conserved between rat and human genes. B: Analysis of plasmids containing intact enhancers. The plots show analysis of transient assays and integrated constructs in two orientations (O1, O2). C: Analysis of plasmids for fine mapping of E2, as in panel B. For each construct, mean and standard deviation of fluorescence intensity were calculated from 3 separate transient transfection experiments, or from multiple integrant clones.

Figure 6. Localization of intronic enhancer E4.

A: Large deletions clones of P+ALB were constructed with restriction enzymes SgfI, XmaI, and EagI (top). Clone ALBΔ2 was then further deleted by PCR (bottom; see Table S2). B: Analysis of the larger constructs by transient transfection. C: Analysis of the larger constructs by integration in two orientations (O1, O2). D: Localization of a strong enhancer toIntron 2 by transient transfection assays. B, D; Mean and standard deviation of fluorescence intensity were calculated from two transient transfection experiments. C; Average and standard deviation of measurements from integrated constructs.

Figure 7. Effects of growth on enhancer function.

A: Selected flow cytometry comparisons. Cells were plated at 25% confluence in 5% fetal calf serum, and allowed to reach confluence (∼72 hr) or switched to 0.1% serum after 24 hr. Exponential growth at 48 hr in serum-containing medium (dotted lines) was compared to growth arrest (solid lines) due to cell confluence at 96 hr (left), or serum starvation at 48 and 96 hr (center and right). A control curve for untransfected cells appears at the left of each panel (solid lines). B: Stimulation of O1 clones by growth arrest. A series of clones systematically resolved the effects onintact Regions 1, 2, ALB, and Enhancers E1 and E2. C: Stimulation ofO2 clones by growth arrest. A more limited series of O2 clones was evaluated as above. B, C; the plots show mean and standard deviation of fluorescence intensity averaged from two separate experiments. Expression was normalized to the level observed for growth at 48 hr in serum.

Figure 8. Sequence analysis of two new Alb enhancers.

Left: E2, localized to a 784 bp region by deletion mapping. Right: the 974 bp minimal region containing E4. Both regions were aligned to the human sequence (DCODE, http://ecrbrowser.dcode.org/); bold, conserved bases; *, insertion in the rat gene; -, insertion in the human gene. Putative transcription factor binding sites were detected with the TRANSFAC database using the RVISTA utility of DCODE, and with binding site models compiled in our laboratory. Sites with high-rank matches by both analyses were annotated in the figure. The ENCODE tracks at the human Alb gene, based on the Human Mar. 2006 (NCBI36/hg18) Genome Assembly (http://genome.ucsc.edu/cgi-bin/hgTracks), include studies of HepG2 cells which mapped chromatin hypersensitive regions and ChIP localizations of HNF4 and C/EBP binding. For both enhancers, the human regions collinear with the illustrated rat sequences have hypersensitive regions and bind HNF4 and C/EBP, although these localizations cannot be precisely mapped to individual binding sites in either genome.