Quantitative characterization of specific genomic promoters using agarose gel electrophoresis

Abstract

Over the past decade a large number of studies have focused attention on the role of nucleosomes as negative and positive regulators of specific nuclear functions. Due to the lack of an analytical method to determine the higher order conformation of the nucleosomal arrays that encompass specific genetic loci (e.g., promoters, enhancers), research emphasis has mostly been centered on chromatin remodeling and histone posttranslational modifications. We have recently developed an agarose gel electrophoresis method that permits us to analyze the higher order structure of specific in vivo assembled chromatin fragments. After calibration using a well-defined in vitro system, we have been able to experimentally determine the size, shape, and conformational flexibility of the Mouse Mammary Tumor Virus long-terminal repeat promoter region in its repressed and activated states. These studies pave the way for widespread analyses of the higher order structure of specific, functionally important chromosomal loci, and in so doing enhance our understanding of the roles that the higher order structure of chromatin play in genome regulations. © 2003 Wiley Periodicals, Inc. Biopolymers 68: 557–562, 2003

RATIONALE

Recent work on gene activation and repression has shifted interest in the field of transcriptional regulation more to the structural aspects of chromatin. An increasing body of evidence suggests that the architecture of the chromatin fiber is strongly linked to its overall role in transcriptional activation or repression.1–4 The commonly depicted transition from an extensively folded state (30 nm fiber), associated with repressed chromatin, to a more open active configuration (also referred to as “bead on the string”) in fact may not be the primary mechanistic switch leading to promoter activation (for review, see5. In contrast, folded higher order chromatin domains may actually be acting to regulate genome function much like folded secondary and tertiary structural motifs mediate protein function.5 The changes in chromatin architecture associated with higher order folding may serve to regulate transcriptional activation by facilitating macromolecular interactions over long distances (up to several thousands of base pairs). In contrast, an unfolded chromatin fiber simply is not capable of mediating the same type of long-distance cis interactions. Data in support of this new model have been provided by recent application of the technique of quantitative agarose gel electrophoresis (QAGE)6, 7 to analyze specific genomic promoters assembled into chromatin in vivo (Georgel et al., submitted). In contrast, previous in vivo analyses of chromatin templates using standard molecular biology approaches such as nuclease sensitivity assays do not yield information about higher order chromatin structure, but rather characterize the primary chromatin structure of the locus being investigated (for reviews, see Refs. 8 and9). Although electron microscopy imaging techniques have been extensively utilized to characterize higher order chromatin structure in vitro,10, 11 they cannot be used to image chromatin samples in an impure environment. Until recently, structural studies of in vivo assembled chromatin by gel electrophoresis have been limited by the inability to analytically characterize specific genomic loci. An equally difficult task has been to determine the protein composition and three-dimensional structure of such in vivo assembled chromatin complexes. Ultimately, no precise measurement of size, composition, or evaluation of chromatin fiber flexibility during promoter activation, within a nucleus or in presence of a crude/nuclear extract, has heretofore been possible. To overcome such technical limitations, we recently adapted the QAGE method to determine the size, flexibility, and higher order structure of specific genomic promoter elements that were excised from their chromosomal location and isolated in unfractionated nuclear extracts.

WHAT IS QUANTITATIVE AGAROSE GEL ELECTROPHORESIS?

Agarose gel electrophoresis is generally thought of as a preparative method for determination of the size of nucleic acid fragments, or for separation of nucleoprotein complexes. However, under specific conditions it also can be used as an analytical tool. The migration of nucleic acids or nucleoprotein complexes is not simply a function of size; charge, shape, and conformational flexibility also are important parameters significantly affecting electrophoretic mobility. To obtain structural information from agarose gel electrophoresis experiments, the QAGE approach initially was used to analyze bacteriophages,10, 12 and subsequently extended to model nucleosomal arrays and chromatin fibers.7, 11, 13 Serwer and colleagues6, 12 empirically determined the relationship between the experimentally measured mobility (μ), the intrinsic gel-free mobility (μ₀′) the effective macromolecular radius (R_e), and the average radius of pores in each running gel (P_e). The equation describing the relationship is as follows:

• (1)

Central to the QAGE approach is the use of multigels, which are composed of 9 or 18 separate running gels embedded in a single 1.5% agarose frame.6, 14 For data analysis purposes, the spherical T3 bacteriophage (R_e = 30.1 nm) is used as an internal standard for the determination of the P_e of each running gel. The μ and μ₀′ of the band(s) of interest are measured experimentally,6, 13 and together with the P_e are used to calculate the R_e using Eq. (1). Multigels allow highly reproducible measurements of the μ₀′ and R_e. The μ₀′ is directly related to the macromolecular surface charge density, while plots of R_e vs P_e yield information about conformational flexibility.14 Previously, it has been difficult to use agarose gel electrophoresis as an analytical technique. However, Eq.(1) establishes a quantifiable relationship that is experimentally accessible through use of agarose multigels.

IN VITRO STUDIES OF CHROMATIN: SETTING THE STAGE

Quantitative agarose gel electrophoresis can be performed under different ionic conditions, allowing characterization of the structure and dynamics of chromatin fibers in vitro (see14. The original set of in vitro experiments using the multigel system were performed with the well characterized 208-12 nucleosomal array model system, which is composed of 12 positioned histone octamers reconstituted onto tandem 208-bp (base pair) repeats of Lytechinus variegatus 5S rDNA.15–17Analysis of these model nucleosomal arrays under low salt conditions demonstrated the reproducibility and utility of the QAGE method for studying chromatin.7 Additionally, the initial experiments found that there are changes in nucleosomal array deformability linked to the level of array saturation, e.g., a single gap in the model nucleosomal array led to pronounced increase in array flexibility in low salt. Of note, the structural changes leading to variations in μ₀′ and R_e also were observed by analytical ultracentrifugation analyses. Subsequent experiments performed in MgCl₂ demonstrated that the R_e provides an accurate measure of salt-dependent chromatin folding,13 and showed that the extent of folding increased with incremental increases in MgCl₂ concentration. Again, the QAGE results were in agreement with those obtained by analytical sedimentation velocity experiments. Taken together, the QAGE results contributed to a model of higher order chromatin folding in which nucleosomal arrays initially condense into an irregular open helical structure mediated by close approach of neighboring nucleosomes, followed by further folding into the classical 30 nm fiber.18 Finally, QAGE experiments were performed to characterize the structural effects of linker histone binding to model nucleosomal arrays.11, 19 These experiments led to the conclusion that linker histone stabilize the intrinsic higher order structures formed by nucleosomal arrays. QAGE analysis of the model chromatin fibers led to the same conclusions as parallel sedimentation velocity experiments, as seen previously with nucleosomal array. Taken together, the studies of nucleosomal arrays and chromatin fibers assembled in vitro from pure components established QAGE as a powerful analytical technique for analysis of chromatin fiber dynamics. We therefore next attempted to surmount a long-standing technical barrier by applying QAGE to the analysis of higher order nucleoprotein structure of specific genomic loci assembled into chromatin in vivo.

APPLICATION OF QAGE TO ANALYSIS OF IN VIVO ASSEMBLED CHROMATIN: CHARACTERIZATION OF CHROMATIN IN UNFRACTIONATED NUCLEAR EXTRACTS

The general approach for analysis of in vivo assembled chromatin is to use restriction enzymes to liberate chromosomal loci into a low salt nuclear extract, followed by electrophoresis of the extract, and detection of specific chromatin fragments by Southern blotting (see below). While performing such experiments (Georgel et al., submitted), we first observed that the analysis of chromatin templates in the presence of unfractionated nuclear extracts by QAGE is complicated by nonspecific binding of extract components to chromatin. Controls performed with 208-12 nucleosomal arrays and DNA in the presence of several different nuclear extracts indicated that a marked supershifted smear is obtained relative to the original DNA or nucleosomal array mobility (Figure 1, lanes 1–3 and lanes 6–8). This behavior potentially compromises the use of crude extracts. Consequently, in an attempt to abolish nonspecific binding, we added an excess of competitor herring sperm (HS) DNA to sequester DNA binding proteins. Addition of 20 μg of HS DNA to 10 μg of nuclear extract prior to addition of 0.2 μg of DNA or model nucleosomal arrays resulted in electrophoretic mobility in the presence of extracts equal to that of the pure DNA or nucleosomal arrays (Figure 1, lanes 4–5 and 9–10). To confirm that the nucleosomal arrays were not dissociated by competitor DNA (via transfer of histones to the HS DNA), two-dimensional agarose gel electrophoresis was performed, and the gel stained with Coomassie blue. A band corresponding to intact 208-12 nucleosomal arrays was observed in the presence of extract and competitor DNA (data not shown), indicating that nucleosomal arrays remain intact under these conditions.

Figure 1. Electrophoretic mobility shift assay of DNA and NA in presence of nuclear extract (NE) and HS DNA (HS DNA). The amount of 0.2 μg 208-12 DNA alone was electrophoresed in 1% agarose (lane 1, see white arrow for DNA mobility). When increasing amounts of NE (respectively 5 and 10 μg) were added, the DNA bands smeared out (lanes 2 and 3). Subsequent addition of 20 μg of HS DNA prevented smearing (lanes 4 and 5). Similar results were observed with 0.2 μg of 208-12 NA. NA alone (lane 6, see black arrow for NA mobility) generated a sharp band. Addition of NE (respectively 5 and 10 μg) resulted in an increasing smear (lanes 7 and 8). HS DNA reversed the trend (lanes 9 and 10).

Analogous multigel experiments were performed using low salt nuclear extracts prepared from yeast, human carcinoma cells and mouse 3134 cells (Figure 2 and data not shown). Results indicated that the μ₀′ and R_e measured in all three extracts in the presence of competitor DNA were the same as those of the parent arrays (Table I), demonstrating that the QAGE is applicable to analysis of chromatin fragments present in crude nuclear extracts. Based on these results, we next performed multigel analyses of specific genomic promoters.

Figure 2. Agarose multigel calibration. Electrophoresis of defined length nucleosomal arrays (NA) in a 0.2–1% agarose multigel in absence (left panel) or presence (right panel) of HS DNA and human carcinoma cell Nuclear Extract (NE) after Southern blotting and hybridization with appropriate end-labeled oligonucleotide (5′ CGC ACC CGT GGC CAG GAC CC 3′). The NA (0.2 μg) are indicated by the black arrows (right side of each panel). Electrophoresis was performed at 1.33 V/cm for 6 h. The horizontal line indicates the origin of electrophoresis. The vertical arrows indicate the direction of electrophoresis and the distance of migrations of individual bands. The μ₀ and R_e extrapolated from the distances of migration are listed in Table I.

Table I. Electrophoretic Properties of 208-12 Nucleosomal Arrays in the Presence of Herring Sperm DNA and Nuclear Extract in Low Salta Buffer

	Re (nm)	−μ0 (× 10^−4 cm^2/Vs)
a 40 mM Tris HCl, 1 mM EDTA, pH 8.0. b HS DNA: Herring Sperm DNA; NE: Dignam Nuclear Extract from yeast or Hey carcinoma cells; h. Carc.: human carcinoma cells.
Nucleosomal arrays	26.6	1.86
Nucleosomal arrays + HS DNA + Yeast NEb	27.07	1.89
Nucleosomal arrays + HS DNA + h. Carc.NEb	26.10	1.83
Nucleosomal arrays + HS DNA + 3134 NEb	27.7	1.90

QAGE OF GENOMIC MOUSE MAMMARY TUMOR VIRUS PROMOTERS

The following section summarizes the results obtained by Georgel et al. (submitted), involving multigel analysis of genomic Mouse Mammary Tumor Virus (MMTV) promoters isolated from mouse 3134 cells.20 These cells contain −200 tandemly repeated copies of the MMTV promoter integrated in chromosome 4.21 The promoter fragments were liberated from the 3134 cell genome by Kpn I restriction, and recovered in an unfractionated low salt nuclear extract. QAGE was performed in presence of appropriate amounts of nonspecific competitor DNA. Band mobilities were determined by hybridization using a probe specific for the genomic MMTV promoter. Controls using model nucleosomal arrays demonstrated that the μ₀′ and R_e determined after SYBR green staining and Southern blotting were the same, validating hybridization as a detection method.

Multigel analysis detected two different MMTV promoter bands present in extracts obtained from cells not exposed to dexamethasone (DEX). The structural parameters obtained from QAGE, together with Western analyses of promoter composition, indicated that one of the bands corresponded to the H1-bound transcriptionally repressed promoter, while the second band was the basally active promoter (note that a heterogeneous mixture of inactive and active promoters was expected because of the 200 integrated copies present in each cell genome). Analogous studies performed with extracts isolated from DEX-treated cells further identified two distinct forms of hormone-activated MMTV promoters. Much to our surprise, these studies also demonstrated that all functional forms of the MMTV promoter could adopt a locally or extensively folded conformation under physiological relevant ionic conditions. These results support a revised model of eukaryotic promoter function in which local interactions between adjacent nucleosomes and their higher order spatial arrangements (i.e., formation secondary chromatin structure) are required for proper function of the MMTV promoter in vivo (see Figure 3).

Figure 3. MMTV promoter model (from Georgel, 20025).The left panel describes the current model for formation of the preinitiation complex on the MMTV-LTR promoter, going from extensively folded to unfolded (beads on a string) leading to an opening of the chromatin fiber. The right panel depicts the alternative model derived from analysis by QAGE. The chromatin fiber remains locally folded and displays flexibility, probably linked to the disruption of nucleosome B.

IMPLICATIONS AND POTENTIAL FUTURE USES OF QAGE

The relevance of our studies of in vivo assembled chromatin using the QAGE method extends far beyond what was determined during the initial investigation of the MMTV promoter in mouse 3134 cells. Perhaps most importantly, the potential applications of QAGE are general, and not limited to studies of one specific genomic locus or to the process of transcriptional activation. Any site of interest that can be cleaved from the chromosome can in principle be studied by QAGE. As such, replication forks, recombination sites, DNA repair sites, and possibly even small nuclear bodies should be equally good candidates for QAGE studies. In the context of transcription, one of the next logical steps will be to study a single copy genomic locus, since the 200 tandemly repeated copies of the MMTV promoter utilized for our initial experiments represent a partially artificial system.

Another area ripe for exploration is the characterization of the structural features of chromatin templates assembled using the Drosophila embryo extract (also referred to as Fly Embryo extract or FEE), initially characterized by Becker and Wu.22 A large number of promoters have been investigated using the FEE and other related extract-dependent chromatin assembly systems.23–25 Although the nucleosome repeat length and overall patterns of nucleosome distributions obtained with the FEE closely resemble that of the native in vivo configuration,23, 26 the use of the crude extract for nucleosome assembly has prevented subsequent characterization of the higher order chromatin structure present under conditions of the functional assays. Thus, QAGE has the potential to determine the structural changes that occur upon sequential addition of the transcription factors or chromatin-associated proteins such as remodeling complexes or histone modifying enzymes in such systems.

In summary, by allowing rigorous analytical studies of specific chromosomal loci and other extract-assembled chromatin templates present in an impure environment, the technique of quantitative agarose gel electrophoresis has enormous potential to enhance our understanding of the roles secondary and tertiary chromatin structure play in genome organization and function.