Moss Tom. DNA-protein interactions: principles and protocols

Подождите немного. Документ загружается.

Osmium modification

129

Fig. 4. In vivo modification of AT tracts: cruciform geometry in bacterial plasmids

as a consequence of physiological salt shock. Bacterial cells containing isogenic plas-

mids with various lengths of AT were salt-shocked and treated with osmium tetroxide

in vivo. Adducts were detected by preparing the DNA, end labeling, and piperidine

cleavage, followed by electrophoresis on thin 6% sequencing gels and autoradiogra-

phy. Strong central modification of the AT tract shows that (AT)

adopts cruciform

geometry in salt-shocked but not in control cells, and that (AT)

and (AT)

but not

(AT)

also adopt cruciform geometry in salt-shocked cells.

c. Once the crystals have dissolved, the reagent is aliquoted ready for use. Glass

containers such as Universals with screw tops should be used, and storage

should be at –70°C. It is best to keep the reagent as several separate aliquots

130 McClellan

and to use them one at a time. One problem with storage is that the glass

containers often crack, which is obviously very dangerous. Volumetric flasks

and non-Pyrex containers are particularly sensitive in this regard. The reagent

should always be thawed with the container inside a glass beaker in the fume

hood. If the reagent is blackish when thawed, discard it; the black color indi-

cates lower oxidation states of osmium, including the metal.

3. By itself, osmium tetroxide is not very reactive with DNA. The species that

attacks DNA is a complex of osmium tetroxide with a heterocyclic compound

such as pyridine or bipyridine (Fig. 5). The attack is on the 5–6 double bond of

thymidines (ref. 3 and references therein). Curiously, neither cytosines nor uracils

show significant modifiability with osmium tetroxide, but guanine residues

occasionally may.

Bipyridine is rather insoluble in water. To make a stock solution, it may be

necessary to heat the bipyridine and water to 80°C for about 30 min. Once

osmium tetroxide is added to pyridine, the mixed reagents should become a straw-

yellow color; if they do not, discard them and obtain fresh osmium tetroxide.

One can change the reactivity of an osmium tetroxide preparation by changing

the concentration of osmium and/or ligand, and the overall degree of reaction can

be varied from single to multiple hits per molecule by changing the time and/or

temperature of the modification reaction. Typical times and temperatures used in

our laboratory to obtain single hits are as follows: 45 min at ice temperature; 15 min

at 20°C; 5 min at 37°C; 1 min at 40°C.

4. Various methods of detecting osmium tetroxide adducts exist, namely retarda-

tion of bands in agarose or acrylamide gels (5), immunoprecipitation (3), cleav-

age by S1 nuclease (5), cleavage by hot piperidine (6), and failure of primer

extension (11). In some cases, inhibition of restriction enzyme cleavage can be

used (10,17).

Fig. 5. Stereochemistry of OsO

attack on thymidines.

Osmium modification

131

5. A good way to do this is to cut with a frequent cutter that

a. has a site near to the end from which it is not required to see signals

b. does not have a site between the end from which it is desired to see signals

and the tract where modifications are expected

For example, the EcoRI site of pXG540 has a HaeIII site 18 bp anticlockwise

(in the direction of the Amp promoter). There is no HaeIII site in the clockwise

170 bp up to the start of the AT tract. So, we often cut and label at EcoRI, and

then do a limit digest with HaeIII. This reults in a long labeled fragment that

has the information we want on it, and a short fragment (18 bp) that we run

off the gel.

6. Osmium tetroxide modification of DNA now has a substantial pedigree, having

been used to modify a wide range of sequences in different unusual conforma-

tions in vitro, and a narrower range of sequences in different conformations in vivo.

Figure 6 shows some of the actual or proposed structures that have been treated

with osmium tetroxide. However, it is very important to be cautious in using

osmium modification results, or any other kind of chemical or enzymatic prob-

ing, to deduce that a particular structure is forming. This is because osmium does

not report on global conformation; it only tells you whether or not a particular T

residue has an exposed 5–6 double bond. If a T is exposed, this could be for

several reasons:

a. The T might be in a single-stranded region of the DNA (e.g., a bubble, mis-

match [18], B–Z junction or cruciform or H-structure loop).

b. The T might be in a helix with a shallow or bulging major groove (e.g., at a

bend or within a GT/AC tract that was forming Z–DNA).

c. The T might be in an overwound helix, with exposure of the 5–6 double bond

resulting from a loss of base overlap, because of the sharp rotation of each

base relative to its neighbors.

These considerations make it very difficult to conclude on the basis of osmium

modification that, for example, Z–DNA is forming; osmium results do not

distinguish in any simple way between what we have called the “U”-structure

(9), Z-DNA and a conformation (possibly the eightfold D-helix), which we have

observed in locally positively stressed AT tracts (19). Detecting the loops and

junctions of cruciform or H-form DNA can sometimes be done in such a way as

to virtually exclude alternative interpretations of the data, but even this is not

always possible; for example, a GC tract with ATAT in the middle will react in

the same way with osmium tetroxide whether it is in a Z-conformation or cruci-

form (11); and there is a published interpretation of chemical modification at

H-forming sequences that is radically at odds with the H-structure model (20).

In addition to these difficulties, there is a need to be cautious about the effect

of the probing chemical on the structural features deduced. Because osmium

modification works under a wide range of conditions, it presents fewer such prob-

lems of interpretation than some other more fastidious chemicals. Nevertheless,

we observe that the ligand used can have an effect on the result obtained; other

things being equal, we find that bipyridine is more likely to give a cruciformlike

132 McClellan

pattern of modification and pyridine is more likely to give a U-structure-like

pattern when superhelically stressed AT tracts are probed in vitro. This probably

has to do with different helix-unstacking potential of the two heterocycles; alter-

natively, it could be an artifact of contaminating cations in the heterocycles or

ion–heterocycle interactions.

Fig. 6. Unusual DNA structures that react with osmium tetroxide.

Osmium modification

133

References

1. Beer, M., Stern, S., Carmalt, D., and Mohlenrich, K. H. (1966) Determination of

base sequence in nucleic acids with the electron microscope. V. The thymine-

specific reactions of osmium tetroxide with deoxyribonucleic acid and its compo-

nents. Biochemistry 5, 2283–2288.

2. Burton, K. and Riley, W. T. (1966) Selective degradation of thymidine and thym-

ine deoxynucleotides. Biochem. J. 98, 70–77.

3. Palecek, E. (1989) Local open DNA structures in vitro and in the cell as detected

by chemical probes, in Highlights of Modern Biochemistry, (Kotyk, A., Skoda, J.,

Paces, V., and Kostka, V., eds.), VSP International Science, Zeist, The Nether-

lands, pp. 53–71.

4. Behrman, E. J. (1988) The chemistry of the interactions of osmium tetroxide with

DNA and proteins, in Symposium on Local Changes in DNA Structure and Their

Biological Implications, Book of Abstracts, p. 6.

5. Buckle, M., Spassky, A., Herbert, M., Lilley, D. M. J., and Buc, H. (1988) Chemi-

cal probing of single stranded regions of DNA formed in complexes between RNA

polymerase and promoters, in Symposium on Local Changes in DNA Structure

and their Biological Implications, Book of Abstracts, p. 11.

6. Lilley, D. M. J. and Palecek, E. (1984) The supercoil-stabilised cruciform of

ColE1 is hyper-reactive to osmium tetroxide. EMBO J. 3, 1187–1192.

7. Johnston, H. and Rich, A. (1985) Chemical probes of DNA conformation: detec-

tion at nucleotide resolution. Cell 42, 713–724.

8. Vojtiskova, M. and Palecek, E. (1987) Unusual protonated structure in the

homopurine. homopyrimidine tract of supercoiled and linearised plasmids

recognised by chemical probes. J. Biomol. Struct. Dyn. 5, 283–296.

9. McClellan, J. A. and Lilley, D. M. J. (1987) A two-state conformational equilib-

rium for alternating A-T)n sequences in negatively supercoiled DNA. J. Mol. Biol.

197, 707–721.

10. Palecek, E., Boublikova, P., and Karlovsky, P. (1987) Osmium tetroxide recognizes

structural distortions at junctions between right- and left-handed DNA in a bacte-

rial cell. Gen. Physiol. Biophys. 6, 593–608.

11. Rahmouni, A. R. and Wells, R. D. (1989) Stabilisation of Z DNA in vivo by

localized supercoiling. Science 246, 358–363.

12. McClellan, J. A., Boublikova, P., Palecek, E., and Lilley, D. M. J. (1990) Super-

helical torsion in cellular DNA responds directly to environmental and genetic

factors. Proc. Natl. Acad. Sci. USA 87, 8373–8377.

17. Palecek, E., Boublikova, P., Galazka, G., and Klysik, J. (1987) Inhibition of

restriction endonuclease cleavage due to site-specific chemical modification of

the B–Z junction in supercoiled DNA. Gen. Physiol. Biophys. 6, 327–341.

16. Duckett, D. R., Murchie, A. I. H., Diekmann, S., von Kitzing, E., Kemper B., and

Lilley, D. M. J. (1988) The structure of the Holliday junction, and its resolution.

Cell 55, 79–89.

13. Dorman, C. J. (1991) DNA supercoiling and environmental regulation of gene

expression in pathogenic bacteria. Inf. Immun. 59, 745–749.

134 McClellan

14. Langermann, S. and Wright, A. (1990) Molecular analysis of the Haemophilus

influenzae type b pilin gene. Mol. Microbiol. 4, 221–230.

15. Horwitz, M. S. Z. and Loeb, L. A. (1988) An E. coli promoter that regulates

transcription by DNA superhelix-induced cruciform extrusion. Science 241,

703–705.

18. Cotton, R. G. H., Rodrigues, N. R., and Campbell, R. D. (1988) Reactivity of

cytosine and thymine in single-base-pair mismatches with hydroxylamine and

osmium tetroxide and its app. lication to the study of mutations. Proc. Natl. Acad.

Sci. USA 85, 4397–4401.

19. McClellan, J. A. and Lilley, D. M. J. (1991) Structural alteration in alternating

adenine–thymine sequences in positively supercoiled DNA. J. Mol. Biol. 219,

145–149.

20. Pulleyblank, D. E., Haniford, D. B., and Morgan, A. R. (1985) A structural basis

for S1 sensitivity of double stranded DNA. Cell 42, 271–280.

Combined SW–Nuclease Footprinting 135

135

From:

Methods in Molecular Biology, vol. 148: DNA–Protein Interactions: Principles and Protocols, 2nd ed.

Edited by: T. Moss © Humana Press Inc., Totowa, NJ

Determination of a Transcription-Factor-Binding

Site by Nuclease Protection Footprinting

onto Southwestern Blots

Athanasios G. Papavassiliou

1. Introduction

The interaction of cell-type-specific or inducible transcription factors with

regulatory DNA sequences in gene promoters or enhancers is a pivotal step in

genetic reprograming during cell proliferation and differentiation and in

response to extracellular stimuli. The study of these interactions and the char-

acterization of the factors involved are, therefore, a critical aspect of gene

control. Transcription factor–DNA interactions in eukaryotes have been dem-

onstrated by a wide variety of biochemical approaches, including deoxyribo-

nuclease I (DNase I) and chemical nuclease footprinting (1–3) (Chapter 3),

methylation protection (4) (Chapter 14), electrophoretic mobility-shift (5,6)

(Chapter 2), and Southwestern (SW) assays (7) (Chapter 17). Despite their

broad applicability, these techniques provide only partial information about

the DNA–protein system under investigation. The first three techniques iden-

tify either the site(s) of transcription factor binding within the DNA (size and

location of nucleotide stretches or atoms on individual bases) or the complex-

ity of the binding pattern (stoichiometry), but do not yield information about

the protein(s) involved. On the other hand, the SW assay reveals the relative

molecular mass of renaturable (on a membrane support) active species in

heterogeneous protein mixtures facilitating their identification, but fails to

localize the exact target element within the probing DNA sequence.

Combining SW and DNase I (but also chemical nuclease and methylation

protection, see below) footprinting methodologies has the dual potential for

accurately determining the size of individual DNA-binding transcription fac-

tors and precisely mapping their cognate binding sites (8) (Fig. 1) . In addition

136 Papavassiliou

to allowing the detection of only fractional binding (footprints in solution can

be obtained only when the binding site[s] is almost completely occupied), cou-

pling in situ DNase I footprinting with the SW technique offers the advantage

of both resolving the protein component and mapping the binding site of com-

plexes formed by either different factors recognizing an identical sequence

within the DNA probe or by two factors interacting in a noncooperative man-

ner with adjacent but distinct sequences. However, this is dependent on the

factor being able to bind to DNA as a monomer or a homodimer. If the active

form is not a single species (i.e., a heterodimeric or heteromeric complex is

required to reconstitute the binding activity), specific DNA binding will not be

detected and the procedure will not be applicable. The most critical stage of the

coupled assay lies within its first part, namely the SW procedure, and concerns

the ability of a transcription factor to efficiently renature into its active form (at

least in terms of DNA-binding capacity) on the membrane. Many transcription

factors are composed of domains with distinct structural conformations, which

aids the process of renaturation. However, because the protein surface immo-

bilized on the membrane poses an impediment on the refolding process, the

Fig. 1. Rationale in a combined SW–DNase I footprinting procedure. The “ham-

mer”-shaped extensions from the DNA-binding protein indicate immobilization on

the blotting membrane. The black-filled sphere indicates the end label in one strand of

the DNA footprint probe. MW denotes the molecular size of the membrane-blotted

DNA-binding protein (estimated by comparing the position of the active membrane

area to the mobilities of coelectroblotted protein MW standards). Arrows mark repre-

sentative sites of DNase I attack. (A) DNase I digestion pattern of free (in solution)

probe; (B) DNase I digestion pattern of protein-complexed (membrane-bound) probe.

Combined SW–Nuclease Footprinting 137

likelihood of successful renaturation increases with increased size of the DNA-

binding transcription factor(s) under study. As a result of the increased kinetic

stability of a membrane-immobilized DNA–protein complex (reversible bind-

ing to even low-affinity proteins is enhanced because excess unbound DNA

has been washed out and hence is not present to compete), reaction parameters

such as the size of the DNA probe, the concentration of DNase I, and digestion

time are no longer determined by the dissociation rate of the complex, a normal

limitation of DNase I protection assays performed in solution.

The fidelity of the combined analytical assay is demonstrated in Fig. 2. Evi-

dently, the structural and functional properties of DNA are not altered by

entrapment on the blotting membrane surface (i.e., the probe exhibits identical

protection pattern and sequence-dependent reactivity with DNase I as it does

in solution). Therefore, this coupled assay provides a fast and reliable method

that will allow the user who has identified specific regulatory regions in the

gene of interest to begin characterizing in detail the transcription factor(s) that

bind to them in a certain cellular milieu.

Modifications of the presented DNase I protection analysis on Southwestern

blotting membranes (in situ) substitute the enzymatic probe for either the

chemical nuclease 1,10-phenanthroline–copper ion (OP–Cu) (9) or dimethyl

sulfate (DMS) (10). Both protocols have the additional advantage of providing

information on the nature (i.e., specific–protected versus nonspecific–

nonprotected) of several membrane-immobilized DNA–protein species often

observed in a SW assay. The molecular and functional properties of DNase I

(i.e., its relatively large size, mode of target searching and binding to cleave,

and requirement for Mg

[which often stabilizes both specific and nonspecific

complexes]) highly reduce its potential to detect these differences. In addition,

these methodologies are useful in rapidly confirming the binding specificity of

a protein isolated by screening a cDNA expression library with recognition-

site DNA.

2. Materials

2.1. SW Blotting

2.1.1. Solutions

1. Sodium dodecyl sulfate (SDS; 20%; w/v): Dissolve 100 g of SDS (Sigma, St.

Louis, MO) in distilled, deionized water to a 0.5 L final volume; heat at 68°C to

assist dissolution (do not autoclave). Store at room temperature in a clear bottle.

A respirator or dust mask should be worn when handling powdered SDS.

2. Lower (separating) gel buffer (4X stock): 1.5 M Tris-HCl, pH 8.8, and 0.4% (w/v)

SDS. Filter through a 0.22-µm membrane filter. Store at 4°C.

3. Upper (stacking) gel buffer (4X stock): 0.5 M Tris-HCl, pH 6.8, and 0.4% (w/v)

SDS. Filter through a 0.22-µm membrane filter. Store at 4°C.

138 Papavassiliou

4. Acrylamide/bis-acrylamide gel mixture: 30% (w/v) acrylamide, 0.8% (w/v) N,N'-

methylene–bis-acrylamide (Bio-Rad, Richmond, CA). Filter through a 0.22-µm

membrane filter. Store at 4°C in the dark. Powdered acrylamide and N,N'-meth-

ylene-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 (work in a chemical

fume hood). 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.

Fig. 2. (A) Schematic outline of the manipulations involved in the combined

SW–DNase I footprinting procedure. In the example presented, a crude lysate of bac-

terial cells overexpressing the proto-oncoprotein c-Jun (a component of the transcrip-

tion factor AP-1) was subjected to a quantitative SW assay utilizing as probe a 134-bp

DNA restriction fragment (5'-end labeled in the coding strand) encompassing the AP-1-

binding sequence of the human collagenase promoter [5'(-72)TGAGTCA3'(–66)]. (B)

DNase I footprinting reactions of the same fragment in solution performed with

increasing amounts (lanes 2–4) of the c-Jun preparation (lane 1, free-DNA probe). The

footprinted region (solid bar) includes in both cases approx 14 bases centered around

the AP-1-binding motif (20).