Moss Tom. DNA-protein interactions: principles and protocols

Подождите немного. Документ загружается.

78 Papavassiliou

digestion pattern of the DNA–protein complexes revealed on the autoradio-

graph with that of free DNA shows a band-free region (footprint) where the

bound protein(s) has prevented access of the enzyme to DNA (see Chapter 3).

In a similar analysis, the DNA is allowed to react mildly with DMS, which

methylates primarily deoxyguanosine residues and renders their phosphodiester

linkages labile under conditions of Maxam–Gilbert chemistry (see Chapter 14).

The binding of a protein(s) to a specific DNA region will result in a protection

of the corresponding bases from chemical modification (2).

The suitability of the above assays in determining the binding sequences of

proteins on DNA is hindered by several disadvantages. First, the clarity of the

footprint is highly dependent on the extent of occupancy of the binding site(s)

(i.e., a “clear” footprint is observed only if all DNA molecules are involved in

complexes). Unfortunately, this is not always easy to achieve, especially when

the concentration and/or purity of the specific binding protein(s) is not satis-

factory. Second, DNA–protein complexes formed in crude extracts may often

be heterogeneous in terms of both binding specificity and kinetic stability.

Therefore, direct footprinting in solution will not correspond to a single spe-

cies, but, instead, reflect an “integral” of the multiple equilibria operating over

the entire region of interest (i.e., the protection pattern will actually represent a

composite of more than one complex, with complexes having a very low disso-

ciation rate dominating the footprint). Finally, two different proteins that

recognize the same sequence within the probe are most likely to yield indistin-

guishable footprints. These drawbacks may be overcome by coupling treat-

ment with a footprinting reagent in solution with the electrophoretic mobility

shift assay (EMSA; also known as gel retardation assay, see Chapter 2) (3–5).

In this approach, the protein and DNA molecules are incubated together, and

the equilibrated reaction mixture is exposed to DNase I or DMS, as before. The

DNA–protein complexes are subsequently isolated from the free probe by elec-

trophoresis in a nondenaturing polyacrylamide gel. Although the negatively

charged free DNA migrates rapidly toward the anode, once it is bound by a

specific protein its mobility decreases (3,4). Following the separation of the

free and bound DNA species, the corresponding bands are cut out of the gel,

and the DNA eluted and analyzed on a sequencing gel. The region(s) of protec-

tion evident in the DNA derived from the complexed fraction, indicates the

binding site (5). Because the complexes are separated from contaminating

unbound DNA fragments, their footprints will be free of background cutting,

and thus considerably more evident. Similar considerations apply when more

than one complex can be formed on the fragment. As long as the DNA-binding

proteins differ in their molecular masses and charges, they will cause altered

electrophoretic mobilities of the corresponding complexes and, hence, differ-

ent migration in the native polyacrylamide gel. These complexes can be iso-

In Gel

OP–Cu Footprinting 79

lated and run in individual lanes on the sequencing gel. Thus, the exposure of

the binding reaction to footprinting reagents, in combination with the fraction-

ation offered by mobility shift gels, permits identification of the regions of

DNA bound by protein in different complexes, even if a low percentage of the

initial DNA molecules has been complexed.

Although one can substantially increase the sensitivity of DNase I or DMS

footprinting experiments in solution by employing the EMSA, several addi-

tional problems have still to be faced:

1. DNase I is a relatively bulky molecule (molecular weight [MW] 30,400) that

cannot cleave the DNA in the immediate vicinity of a bound protein because of

steric hindrance. As a result, the region(s) protected from cutting extends beyond

the actual protein-binding site.

2. The nonrandom nature of DNA cleavage by DNase I makes it impossible to assess

the involvement in protein binding of nucleotides that lie in an area of the frag-

ment resistant to the endonucleolytic activity of this enzyme (e.g., tracts of A and

T residues, or TpA [as opposed to ApT] dinucleotide islands scattered within or

adjacent to the binding site), so that binding sites or parts of binding sites may not

be detected.

3. The primary site of reaction of DMS with B-DNA is the N-7 atoms of guanine

bases, which are located in the major groove. Thus, those guanines in close prox-

imity with the protein will be protected from methylation. However, if a protein

primarily makes contacts with a DNA sequence in the minor groove, or if there

are no guanine residues in a major groove-binding site, DMS will not reveal

these interactions.

4. In many instances, particularly when a complex has a relatively high “off” rate,

the bound protein can dissociate from the protected DNA fragment and reassoci-

ate to other DNA fragments that have already been nicked by DNase I or modi-

fied by DMS. In this case, the DNA-cleavage pattern derived from the complexed

fraction will closely resemble that of the uncomplexed DNA, rendering it diffi-

cult to observe a footprint. The limitations imposed by the size and the sequence

or base specificity of the aforementioned footprinting reagents, as well as the

problem of protein exchange from the binding site(s) during treatment, are cir-

cumvented by merging the advantages inherent in the EMSA, with the subse-

quent exposure of the gel (hence of the resolved complexes while embedded in

the polyacrylamide matrix) to a chemical DNA-scission reagent namely the 1,10-

phenanthroline–copper ion (OP–Cu) (6).

1.1. OP–Cu as a Footprinting Agent

1.1.1. Chemistry of DNA Cleavage

1,10 Phenanthroline–copper (Fig. 1) is an efficient chemical nuclease that

cleaves the phosphodiester backbone of nucleic acids at physiological pH and

temperature by oxidation of the deoxyribose (DNA) or ribose (RNA) moiety

80 Papavassiliou

(7). The kinetic scheme of the reaction is summarized in Fig. 2. The first step is

the formation of the 1,10-phenanthroline-cupric ion coordination complex,

under conditions that favor the 2:1 stoichiometry ([OP]

). The DNA-

scission process is initiated by adding a reducing agent, usually 3-mercapto-

propionic acid (a thiol), to the aerobic reaction mixture containing the target

DNA. Under these conditions, the 2:1 cupric complex is reduced to the 2:1

cuprous complex ([OP]

) that is, in turn, oxidized by molecular oxygen to

generate hydrogen peroxide. Hydrogen peroxide is an essential coreactant for

the chemical nuclease activity and can be generated as described above or

added exogenously (8). The tetrahedral cuprous complex, present at the steady-

state concentration defined by the experimental conditions (note the feedback

mechanism in Fig. 2), then binds reversibly to the minor groove of DNA to

form a central intermediate through which the reaction is funneled (9). The

DNA-bound cuprous complex undergoes in situ a one-electron oxidation by

hydrogen peroxide to form a short-lived, highly reactive DNA-bound copper-

oxo species that can be written either as a hydroxyl radical coordinated to the

cupric ion or as a copper-oxene structure (Fig. 2). This species then attacks the

H1'-deoxyribose protons of nucleotides, which are accessible in the minor

groove; this reaction initiates a series of reactions culminating in cleavage of

the phosphodiester backbone (9). Reaction rates at any given sequence posi-

tion depend on the stability of the intermediate formed between DNA and

(OP)

and on the orientation and proximity of the copper-oxo species rela-

tive to the C1'-deoxyribose hydrogen in the minor groove. Because both crite-

ria are met satisfactorily in B-DNA sequences, the tetrahedral cuprous complex

prefers B-DNA as its substrate. Such stereoelectronic interactions are less effi-

cient in the broad minor groove of A-DNA and not possible in Z-DNA, which

Fig. 1. Structure of 1,10-phenanthroline complexed with copper(I) ion (OP–Cu).

In Gel

OP–Cu Footprinting 81

Fig. 2. Schematic representation of the kinetic mechanism for the nuclease activity of 1,10-phenanthroline

–copper ion.

82 Papavassiliou

has practically no minor groove; as a result, A-DNA is cleaved at 25–33% of

the rate with which B-DNA is cleaved and Z-DNA is not cleaved at all (9).

The products of the strand-scission event include the free base, DNA frag-

ments bearing 5'- and 3'-phosphorylated termini, and the deoxyribose oxida-

tion product 5-methylene-2-furanone (10). The DNA-chain cleavage

reaction can be efficiently quenched by adding to the mixture 2,9-dimethyl-

1,10-phenanthroline (2,9-dimethyl–OP). This phenanthroline derivative can

also chelate copper ions to form a minor groove-associated cuprous complex

(thus competing with [OP]

), but the reduction potential of the Cu

/Cu

couple is too positive to allow significant nuclease activity under normal assay

conditions (11).

1.1.2. OP–Cu Footprinting Following EMSAs

In as much as the structural and functional properties of DNA are not altered

by entrapment in a polyacrylamide gel matrix (6), the small size and the ready

diffusibility of all reaction components in solid supports permit the coupling of

OP–Cu footprinting with the EMSA to study DNA–protein interactions

(12,13). In this method, the DNA-binding reaction is performed as usual, elec-

trophoresed under established, nondenaturing conditions, and the entire mobil-

ity shift gel is immersed in a footprinting reaction mixture containing

1,10-phenanthroline, cupric ion, and 3-mercaptopropionic acid. Following the

reaction quench with 2,9-dimethyl–OP, footprints are obtained after elution of

the radioactive free and protein-bound DNA cleavage products from the

mobility shift gel and analysis on a sequencing gel (Fig. 3). Because the

nuclease activity of (OP)

produces 3'-phosphorylated and 5'-phosphory-

lated ends as cleavage products, sequencing gels can be accurately calibrated

with the Maxam–Gilbert sequencing reactions.

1.2. Advantages of OP–Cu over Other Footprinting Agents

1.2.1. General Considerations

The nuclease activity of (OP)

bears several advantages as a footprinting

reagent relative to protection analyses using DNase I or DMS as a probe. First,

the (OP)

chelate is a small molecule (compared to DNase I) that permits

cleavage closer to the edge of the DNA sequence protected by protein binding

and, therefore, a more precise definition of it. Second, because the scission

chemistry involves attack on the deoxyribose moiety, (OP)

is able to cut at

all sequence positions regardless of base. However, the intensity of cutting

(rate of cleavage) does depend on local sequence, with attack at adenines of

TAT triplets being most preferred (14; see also legend to Fig. 3). Interestingly,

a preference for C-3',5'-G steps, rather than T-3',5'-A steps, is observed at a

In Gel

OP–Cu Footprinting 83

phenanthroline to copper ratio of 1:1, which strongly favors formation of the

OPCu

complex (15). Nevertheless, the cutting patterns obtained with

(OP)

are usually sufficiently well-defined to identify protected regions,

even though this endonucleolytic agent exhibits some degree of sequence speci-

ficity in its rate of cleavage of naked DNA. Third, because (OP)

binds to

the minor groove of DNA, it will reveal minor-groove interactions. Because

the binding of the coordination complex should be restricted to three base

pairs, the complex is more sensitive to local, protein-induced conformational

changes than DNase I, which by possessing an extended minor groove-bind-

ing site, may be unable to sense. In this context, the complex will also detect

binding in the major groove when its approach to its minor groove-binding site

is sterically blocked or if the interaction of the protein in the major groove

alters the minor groove geometry so that the tetrahedral coordination complex

binds poorly (both being frequent features of DNA–protein interactions). Fur-

thermore, because of the difference in their respective mechanisms of cleav-

age, DNase I and (OP)

probe different aspects of the structure of a

DNA–protein complex. DNase I cleavage relies on the accessibility of a par-

ticular phosphodiester bond, and thus protection is indicative of an interaction

on the outer face of the DNA helix. In contrast, protection from (OP)

mediated cleavage is most likely caused by the inhibition of its binding to the

minor groove and implies that a portion of the protein occupies at least the

minor groove. Finally, in contrast to other chemical nucleases such as ferrous

EDTA (introduces single-stranded nicks in DNA through the generation of

diffusible hydroxyl radicals; see Chapter 5), the nucleolytic activity of OP–Cu

is not inhibited by glycerol, a free radical scavenger, which is present in most

protein storage buffers.

1.2.2. Benefits of OP–Cu Footprinting Within Mobility Shift Gels

The major advantage of the combined OP–Cu footprinting procedure arises

from the topography of treatment: Preformed DNA–protein complexes are

exposed to the chemical nuclease within the gel (i.e., not prior but subsequent

to an electrophoretic mobility shift experiment). This characteristic of the tech-

nique makes it ideal for protection analysis of kinetically labile complexes (16).

At least three factors account for the latter. The first is that the background

cleavage is greatly reduced by the separation of unbound DNA from the DNA–

protein complex(es) pool. The second factor is the so-called “caging effect”

(3,4). The gel matrix forms “cagelike” compartments that prevent a dissoci-

ated protein from diffusing away from the DNA, so that by enhancing

reassociation, the apparent affinity constant will be higher than the true value.

The protein could also interact with the gel matrix, thereby orienting its diffu-

sion toward reassociation. Whatever the mechanism(s), the increase in stabil-

84 Papavassiliou

In Gel

OP–Cu Footprinting 85

ity of the complex contributed by the gel leads to a more efficient blockage of

the access of the (OP)

chelate to the protein-binding DNA segment.

The third factor comes from the nature and site of action of the cupryl

intermediate through which the reaction is funneled, and it acts synergisti-

cally with the previous one. Because this highly reactive oxidative species is

generated near the surface of the DNA (in situ), diffusible radicals, if formed at

all, will have a short or restricted diffusive path and, therefore, will be unable

to achieve a fast equilibrium distribution along the DNA polymer. Conse-

quently, protein-binding sites exposed during multiple dissociation events will

escape the nucleolytic attack most of the time and hence remain intact.

In addition to the fact that discrete complexes with defined stoichiometries

and a wide range of kinetic stabilities can be mapped simultaneously, the in

situ OP–Cu footprinting procedure is superior to oligonucleotide-binding com-

petition assays in the analysis of multiple complexes frequently obtained in

electrophoretic mobility shift experiments employing unfractionated extract

preparations. For example, multiple retarded bands can arise from protein–

protein interactions between a non-DNA-binding transcription factor(s) and a

specific DNA-binding protein, or from two proteins binding to distinct DNA

sequences in a cooperative manner (12). Although both complexes would be

Fig. 3. (previous page) Outline of the combined electrophoretic mobility-shift/in

gel OP–Cu footprinting assay. (A) DNA restriction fragments containing a protein-

binding site(s) are labeled with

P at a unique end and incubated with a crude or

partially purified extract containing the DNA-binding protein(s) of interest, under

optimized binding conditions. (B) After equilibration of the DNA-binding reaction,

the free and bound DNA fragment populations are separated by electrophoresis through

a nondenaturing polyacrylamide gel; the gel is then transferred into a buffer-contain-

ing Pyrex dish, and the retarded and unretarded DNA species are exposed in situ to the

nuclease activity of (OP)

. The two DNA fractions are subsequently located by

autoradiography of the wet gel, excised and eluted from the gel matrix, precipitated,

and recovered in formamide buffer. (C) Samples are heat-denatured and equal amounts

of radioactivity from the two fractions are electrophoresed on a denaturing polyacryla-

mide gel (DNA sequencing gel) and autoradiographed. In the sample prepared from

the free-DNA band, bands will appear in the gel corresponding to positions of protein

binding. For the sample(s) prepared from the protein–DNA band(s), bands will appear

at all positions except those bound by the protein(s) (protected region). The particular

example depicts the OP–Cumapping of a DNA–protein complex formed between bac-

terially expressed LFB1 (a liver-specific transcription factor) and an oligonucleotide

bearing its binding site within the –95 to –54 region of the α1-antitrypsin promoter.

Arrowheads connected by line demarcate the footprinted site. The enhanced cleavage

observed within the protein-binding site in the free-DNA sample is the result of the

presence of repeated TA elements in this sequence (see Subheading 1.2.1.).

86 Papavassiliou

abolished by competition with oligonucleotides, these possibilities can be

readily distinguished by direct footprinting within the gel.

1.3. Additional Applications and Outlook

The nucleolytic activity of (OP)

in a polyacrylamide matrix has been

also demonstrated to be a viable means of gaining insight into the interactions

of RNA-binding proteins with their recognition sequences (17,18). Applica-

tion of OP–Cu in this context may be invaluable toward defining structural

perturbations in RNA on protein binding and mapping the binding domains of

various proteins. Because of the preferential nucleolytic activity of (OP)

toward single-stranded bulge and loop RNA regions (double-stranded stem

regions can be cut at elevated concentrations of the chemical nuclease), hyper-

sensitive sites may be obtained on footprinting an RNA–protein complex

following a gel retardation assay. Such sites would imply an unwinding of a

helical structure on protein binding or perturbations in the minor groove acces-

sibility of the bound RNA molecule.

The in gel OP–Cu footprinting methodology has already expanded the

“tool box” available to investigators wishing to explore the structure and

function relationships of nucleic acid–protein complexes, and emerging

improvements in the chemical mechanism (e.g., DNA-strand scission by the

coordination complex of OP with a non-redox-active metal) as well as future

modifications will likely make this technology even more efficient and

broadly useful.

2. Materials

2.1. Analytical and Preparative EMSA

2.1.1. Solutions

1. A variety of binding and gel buffers are commonly employed in EMSA (see Chap-

ter 2). A suitable binding buffer is 10 mM HEPES pH 7.9, 10% glycerol,

0.1 mM EDTA, 0.5 mM tetrasodium pyrophosphate, and 0.5 mM PMSF. The

most common gel buffers are Tris-glycine: 50 mM Tris, 2.5 mM EDTA, and 0.4 M

glycine; 0.5X TBE: 45 mM Tris, 45 mM boric acid, and 1 mM EDTA; Tris–acetate:

6.7 mM Tris-HCl, pH 7.5, 3.3 mM sodium acetate, and 1 mM EDTA.

2. Ammonium persulfate (10%; w/v): Weigh out 1 g of ammonium persulfate and

put it in a sterile plastic tube containing 10 mL of distilled, deionized water.

Vortex vigorously until the salt is completely dissolved. Filter through a 0.22-µm

membrane filter. This solution may be stable for a period of a few days at 4°C,

but it is recommended that you prepare it freshly for each new gel. Ammonium

persulfate is extremely destructive to tissue of the mucous membranes and upper

respiratory tract, eyes, and skin. Inhalation may be fatal. Exposure can cause

gastrointestinal disturbances and dermatitis. Wear gloves, safety glasses, respira-

In Gel

OP–Cu Footprinting 87

tor, and other protective clothing and work in a chemical fume hood. Wash thor-

oughly after handling.

3. Dye-containing binding buffer: 0.05% (w/v) bromophenol blue in 1X optimized

binding buffer (store at 4°C after filtering).

2.1.2. Reagents/Special Equipment

1. Highly purified duplex DNA fragment labeled exclusively at one of its four ends

(5' or 3'); use standard procedures for unique labeling (19). All necessary precau-

tions should be observed to minimize exposure to ionizing radiation during label-

ing and isolation of the probe. Consult the institutional environmental health and

safety office for further guidance in the appropriate use of radioactive materials.

2. Reagents employed in the optimized binding reaction.

3. (16–18) × (16–18)-cm front and back glass gel electrophoresis plates: The plates

must be absolutely clean before use. Wash them with warm soapy water; then,

holding them by the edges, rinse several times first in tap water and then in deion-

ized water. Finally, rinse with ethanol and let them air-dry. Using a pad of

Kimwipes, siliconize the inner side of the back plate with a 2% dimethyl-

dichlorosilane solution in 1,1,1-trichloroethane in a chemical fume hood (this

product is particularly toxic; gloves, safety glasses, respirator, and other protec-

tive clothing should be worn when handling it.

4. 0.3-cm spacers.

5. Electroresistant plastic tape (e.g., 3M yellow electrical tape).

6. 0.22 and 0.45-µm filters (Millipore, Bedford, MA).

7. N,N,N',N'-Tetramethylethylenediamine (TEMED; Bio-Rad, Richmond, CA).

TEMED is extremely destructive to tissue of the mucous membranes and upper

respiratory tract, eyes, and skin. Inhalation may be fatal. Prolonged contact can

cause severe irritation or burns. Wear gloves, safety glasses, respirator, and other

protective clothing and work in a chemical fume hood (TEMED is also flammable!).

Wash thoroughly after handling.

8. 3-mm gel comb with 10-mm-wide teeth.

9. 10-mL syringe and 18-gage needle.

10. 100- to 200-µL Hamilton syringe.

11. Additional reagents and equipment: Powdered acrylamide and N,N'-methylene–

bis-acrylamide (Bio-Rad); plenty of binder clamps (fold-back spring clips); razor

blades; polyacrylamide gel electrophoresis apparatus; constant current power

supply; peristaltic pump for recirculating electrophoresis buffer (if required);

siliconized 1.5-mL Eppendorf microcentrifuge tubes; spatula. Acrylamide and

N,N'-methylene–bis-acrylamide are potent neurotoxins and are absorbed

through the skin. Their effects are cumulative. Wear gloves and a face mask

when weighing these substances and when handling solutions containing them.

Although polyacrylamide is considered to be nontoxic, it should be handled

with care because of the possibility that it might contain small quantities of

unpolymerized acrylamide.