Moss Tom. DNA-protein interactions: principles and protocols

Подождите немного. Документ загружается.

584 Gray

10 min, carbon side up, on a filter paper that is under the lamp and approx 12 cm

from the bulb.

3.2. Specimen Tilting

Although specimen grids are readily tilted in an electron microscope fitted

with a goniometer, when dealing with flexible helical nucleoprotein complexes

one is confronted with two problems: first, to find a specimen that will show

the helical asymmetries when it is tilted, and second, to determine which end

of a tilted helix is closer to the viewer when the helix is seen on the fluorescent

screen, in negatives, and in prints. We deal with these problems in the following:

1. Flexible nucleoprotein helices will generally not have straight helical axes and

uniform turns as in Fig. 2, but a much less regular structure. It is essential to

select a segment of a helical complex in which the turns are as regular as can be

found, with a roughly linear (or only gently curved) helix axis extending for 5–10

helical turns. Figure 3 contains a set of drawings taken from an actual tilt experi-

ment (3). The helix axis of a relatively linear and regular helical segment has

been oriented approximately perpendicular to the tilt axis (which is horizontal in

Fig. 3) to maximize the changes observed upon tilting. The characteristic left-

side and right-side indentations are seen when the complex in Fig. 3 is tilted

in opposite directions (+55°, –55°), even though there is significant flexing in

the helix axis. The change of indentations from one side to the other is faintly

visible even in a diagonally oriented helical segment (arrows), but the effect is

more convincing in a segment that is perpendicular to the tilt axis.

2. The complexes must also have been prepared under conditions that yield a mini-

mum of flattening of the helical structure; the use of a relatively deep layer of

negative stain will tend to help support the three-dimensional structure. Flatten-

ing of the helix makes the left-hand-side and right-hand-side indentations in tilted

helices more difficult to observe, but they still can be seen when the helix is

distorted. Note that in Fig. 3 the width at a appears to be about half of the width

at b in the helix that is tilted –55°, indicating that the helix is flattened so that the

height of a helical coil above the support film is roughly half of the width of the

coil measured in a direction parallel to the support film.

3. In order to determine the absolute orientations of helical specimens as they are

seen on the fluorescent screen of an electron microscope, focusing effects are

utilized. An asymmetric marker (such as a macroscopic letter R punched on a

specimen grid) can be used to correlate what is seen on the screen with the known

position of the tilted grid (oriented, for example, so that the top edge of the R is

uppermost in the column at a tilt angle of +55°). We used such a device in the

top-entry stage of our Zeiss EM10C to demonstrate that when the objective lens

current is adjusted to focus on the central portion of a steeply tilted grid, then the

edge of the grid that is uppermost in the microscope and furthest from the upper

polepiece of the objective (imaging) lens will be overfocused (lens current too

strong to focus on it), whereas the lower edge of the grid will be underfocused

Electron Microscopy of Nucleoprotein Complexes 585

(lens current too weak). These focusing effects can be used to determine which

end of any helical complex is uppermost in the column; that same upper end of

the helix will be, in effect, the end that is nearest to the viewer who is looking at

the helix on the fluorescent screen.

4. The differing Fresnel patterns due to underfocusing and overfocusing of a tilted

specimen support film are visible in a negative bearing an image of the specimen.

Hence, proof of the three-dimensional orientation of the specimens is contained

in each negative. To correctly interpret the hand of a helix from an image, the

negatives must be viewed from the emulsion side, which in the microscope faces

Fig. 3. Drawings made from electron micrographs of a left-handed helical complex

of the IKe gene 5 protein with single-stranded fd viral DNA. The center image is of

the nontilted complex that lies in the plane of the page. The upper image is of the same

complex tilted –55° around a horizontal axis, so that the top of the complex is below

the plane of the page and the bottom of the complex is closer to the observer. The

lower image is of the same complex tilted +55°, so that the bottom of the complex is

below the plane of the page and the top of the complex is nearer to the observer. The

drawings show the helical coils as three-dimensional structures to show how they can

account for the projection images seen in the original electron micrographs, which are

shown in ref. 3. (Reprinted with permission from ref. 3.)

586 Gray

in the same direction (toward the electron beam and toward the viewer) as does

the viewing surface of the fluorescent screen. It is, of course, essential that the

operator correctly recognize the patterns corresponding to overfocusing as

opposed to underfocusing; if in doubt, the patterns can be demonstrated by mak-

ing exposures of deliberately underfocused and overfocused specimens at zero

degrees of tilting.

5. The Fresnel patterns will also be visible in prints made at a suitable magnifi-

cation. To present the helical asymmetries correctly, the prints must be made

with the emulsion side of the negative facing the enlarger light source, so that

the print will represent the same view of the object as that which is seen on the

fluorescent screen.

6. Finally, we emphasize that it is essential to demonstrate that the indentations

between helical turns switch from one side to the other of the same helical nucle-

oprotein segment as the helix is tilted in opposite orientations. We find that the

flexing and partial compression of nucleoprotein helices can produce structures

that show indentations on one side in one tilt orientation, but that do not show

indentations on the other side when the helix is tilted in the opposite orientation.

A satisfactory proof of the helical symmetry (hand of the helix) requires that both

symmetrically related effects be demonstrated in the same helical segment.

4. Notes

1. The purity of the water used as solvent can be critical to the success of prepara-

tions for electron microscopy. The chemical content of the water supplied to a

laboratory varies greatly with the locale, and it sometimes happens that proce-

dures that worked in one location will fail in another, when the only reagent not

carried to the new location is the water. If difficulties are encountered with

experimental procedures, alternative sources of water and alternative water puri-

fication protocols should be explored.

2. The exercise of care to preclude the contamination of solutions is essential to the

success of these procedures. Negative staining of specimens mounted on glow-

discharge-activated grids is a widely used technique, but long and flexible nucle-

oprotein helices are particularly susceptible to the effects of contaminants that

can interfere with the uniform adsorption of the complexes to an activated carbon

film. The adsorption may fail to give satisfactory results if appropriate precau-

tions are not taken.

3. We find that a more uniform layer of uranyl acetate stain can be obtained if the

carbon-coated grids are hydrated prior to step 5 of Subheading 3.1. This is done

by placing the grids carbon side up in a small plastic Petri dish, 2 cm in diameter,

and then floating the small Petri dish on distilled water that has been brought to

90°C and poured into a standard Petri dish bottom. Another standard Petri dish

bottom is inverted over the first, creating a chamber in which steam is captured

and will be exposed to the carbon coatings on the grids. After 10 min of hydra-

tion, the grids are dried under a lamp for 10 min and are immediately subjected to

glow-discharge activation. The depth of the uranyl acetate stain layer can be con-

Electron Microscopy of Nucleoprotein Complexes 587

trolled by varying the times of hydration exposure and drying. In our hands, this

procedure yields satisfactory staining on every grid, provided that there is no

interference resulting from contaminants in the specimen.

Acknowledgments

The author gratefully acknowledges support of this work by National Insti-

tutes of Health Research grant 5-RO1-GM34293-03 (to C.W.G.), by NIH

Biomedical Research Support grant 2S07-RR07133-21, by NIH Small Instru-

mentation grant 1-S15-NS25421-01, and by National Science Foundation

Instrumentation grant PCM-8116109.

References

1. Bujalowski, W. and Lohman, T. M. (1991) Monomer–tetramer equilibrium of

the Escherichia coli ssb-1 single-strand binding protein. J. Biol. Chem. 266,

1616–1626.

2. Kodadek, T. (1990) The role of the bacteriophage T4 gene 32 protein in homolo-

gous pairing. J. Biol. Chem. 265, 20,966–20,969.

3. Gray, C. W. (1989) Three-dimensional structure of complexes of single-stranded

DNA binding proteins with DNA: IKe and fd gene 5 proteins form left-handed

helices with single-stranded DNA. J. Mol. Biol. 208, 57–64.

4. Bochkarev, A., Pfuetzner, R. A., Edwards, A. M., and Frappier, L. (1007). Struc-

ture of the single-stranded-DNA-binding domain of replication protein A bound

to DNA. Nature 385, 176–181.

5. Skinner, M. M., Zhang, H., Leschnitzer, D. H., Guan, Y., Bellamy, H., Sweet, R.

M., et al. (1994). Structure of the gene V protein of bacteriophge f1 determined by

multiwavelength X-ray diffraction on the selenomethionyl protein. Proc. Natl.

Acad. Sci. USA 91, 2071–2075.

6. Olah, G. A., Gray, D. M., Gray, C. W., Kergil, D. L., Sosnick, T. R., Mark, B. L.,

et al. (1995) Structures of fd gene 5 protein-nucleic acid complexes: a combined

solution scattering and electron microscopy study. J. Mol. Biol. 249, 576–594.

7. Gray, C. W., Brown, R. S., and Marvin, D. A. (1981) Adsorption complex of

filamentous fd virus. J. Mol. Biol. 146, 621–627.

8. Finch, J. T. (1972). The hand of the helix of tobacco mosaic virus. J. Mol. Biol.

66, 291–294.

9. Haschemeyer, R. H. and Myers, R. J. (1972). Negative staining, in Principles and

Techniques of Electron Microscopy (Hayat, M. A., ed.), vol. 2, Van Nostrand

Reinhold, New York, pp. 101–147.

10. Valentine, R. C., Shapiro, B. M., and Stadtman, E. R. (1968). Regulation of

glutamine synthetase, XII. Electron microscopy of the enzyme from Escherichia

coli. Biochemistry 7, 2143–2152.

STEM of DNA–Protein Complexes 589

589

From:

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

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

Scanning Transmission Electron Microscopy

of DNA–Protein Complexes

Joseph S. Wall and Martha N. Simon

1. Introduction

The scanning transmission electron microscope (STEM) at Brookhaven

National Laboratory* (BNL) is nearly unique in its ability not only to image

isolated unstained and unshadowed biological molecules but also to obtain

quantitative information about them such as their oligomeric state (1) (see

Note 1). This makes it ideal, in principle, for looking at protein–DNA com-

plexes and obtaining information about the masses bound. However, there are

very stringent requirements on the purity and stability of the samples.

The quantitative STEM offers significant advantages over other techniques

for characterizing biological complexes (see Note 2). The STEM operates in a

dark-field imaging mode which gives high contrast for small objects. The

STEM forms an image one point at a time by rastering a finely focused elec-

tron beam over a specimen in a TV-type scan. Typically 1000 electrons are

used for the readout of each picture element (pixel) in the specimen. Most of

these pass directly through the specimen without losing energy or changing

direction. However, a few of these electrons interact with the specimen and are

scattered into two annular detectors. The number of scattered electrons is

directly proportional to the mass thickness at that point.

The digital STEM image is a two-dimensional projection map of the

specimen’s mass distribution. Adding up the mass in regions of the image

containing a complex gives the molecular weight, which usually identifies

the complex and determines its oligomeric state. Thus, a link is established

between biochemistry and imaging.

*The BNL STEM is a NIH Supported Resource Center, NIH P41-RR01777, with additional

support provided by the Department of Enrgy and Office of Biological and Environmental Research.

590 Wall and Simon

The high contrast and signal-to-noise-ratio (S/N) of a STEM dark-field

image permits direct visualization of unstained double-stranded DNA (dsDNA)

and can provide an accurate measurement of its total length and the position of

a bound complex (see Note 3). The high contrast of the STEM also reveals any

problems with the specimen such as denatured protein or residual salt. This

forces rigorous attention to details of specimen preparation that, although some-

times annoying, leads to results having high credibility.

We distinguish three types of complex that place different constraints on

specimen preparation, imaging, and analysis: (1) one-dimensional complexes

(2) such as a protein binding to a specific sequence of nucleic acid, (2) two-

dimensional complexes (3) such as chromatin, filamentous viruses and recA-

type filaments, and (3) three-dimensional objects such as ribosomes (4),

“spherical” virus particles (5),and intermediates in such viral assembly.

Objects of the last type tend to be the easiest to work with because the com-

ponents are bound together in a relatively rigid and stable complex. Adsorbing

such objects from solution onto a suitable EM substrate such as carbon film

generally causes minimal perturbation. The particles tend to have characteris-

tic shapes and relatively few distinct orientations, so it is easy to recognize

broken or incompletely assembled particles. Compact shapes with distinct bound-

aries make it relatively easy to identify complete particles for mass measurements.

Two-dimensional complexes are usually curved and may also be tangled in

solution, so depositing them on a planar substrate may cause distortions or

topological problems. In this case also, the outside edges of the complex are

usually composed of a relatively uniform compact protein shell, which is not

significantly disrupted by contact with the substrate. Any areas that have

unraveled are easily recognized as different from sharp, well-preserved sec-

tions for mass/length determinations.

One-dimensional complexes tend to be the most difficult with which to

work. In solution, the dsDNA itself tends to be a random coil with low rigidity,

so docking onto the substrate can give an uninterpretable tangle if the strands

are long. Searching for bound proteins on this can be slow, so the best approach

is to use the shortest strands compatible with the experiment. A more serious

problem comes from the fact that such complexes contain relatively large

stretches of bare nucleic acid that attach poorly to the carbon film substrate.

The entire coil in solution is undergoing rapid translational and rotational

Brownian motion. If the protein sticks first, as usually happens, the mechanical

forces transmitted to the binding region are frequently greater than the specific

binding forces and the complexes are disrupted. Shorter strands help, but cova-

lent crosslinking is usually required to give reproducible complexes. Alterna-

tively, poly-lysine pretreated grids will bind the dsDNA well, but, sometimes,

this forces a weaker binding protein off so that the complexes are disrupted,

STEM of DNA–Protein Complexes 591

and, again crosslinking may be required. However, if the strands are too short

(e.g., made from oligomers of 20 bases) such that the protein complex covers

them, it is not possible to determine the orientation of the complex on the DNA.

The use of general fixatives to crosslink, such as glutaraldehyde and formal-

dehyde, raises the question of generating artifacts such as nonspecific binding

or binding of extraneous proteins and consequent increases in masses. The best

scenario is to have is a specific crosslinker between the protein(s) and the

dsDNA, but that is not always available.

Whatever is used should be tested on the complex, and the minimum condi-

tions to maintain it should be determined. This can be done by footprinting, gel

retardation, or some other assay of the crosslinking conditions. The cross-

linking need not be 100% and that should actually probably be avoided, because

anything that is overcrosslinked will be uninterpretable (See Note 4.) If more

than 50% of the molecules are in complexes and the concentration is good

(roughly 10 complexes per square micron), searching in the microscope can be

fairly rapid if the strands are reasonably short (<1000 bp). The histogram of

mass bound as a function of distance from the end of the strand is the final test

of specimen quality.

2. Materials

We will describe in some detail the materials we use, as the physical purity

of most materials is critical to the success of a STEM project. The grids for the

STEM are prepared by the STEM staff, but the specimens themselves come

from the users. (See Note 5.) Knowing the details involved in specimen prepa-

ration enables users to understand the problems that can arise. One of the most

important is the physical purity of all of the materials involving the sample,

which may end up on the grid in the STEM.

2.1. Water, Buffers, and Salts

1. Of surprising importance is that the water used in all the steps of sample prepara-

tion must be physically clean. The water in some labs is very impure. STEM

water is deionized and freshly distilled daily. It is used for all grid and sample

preparation as well as for all buffers (see Note 6).

2. All solutions must be physically clean. Anything “dirty” involved in the purifica-

tion such as fragments of column or gel material will show up on the grid and

interfere with the observations and analysis. Any physical additives that scatter

electrons will also interfere (see Note 7).

3. Most samples can be applied to the grid in the buffer which is necessary for their

biological activity. Additives such as sucrose, glycerol, and dithiothreitol will

usually wash off. If necessary, a sample can be applied to the grid in high salt.

(A few washes with a high concentration of ammonium acetate will usually

remove it.) (See Note 8.) However, some buffers and salts are known to cause

592 Wall and Simon

problems (see Note 9). HEPES and MOPS are good buffers. The final washes of

the grid before freeze-drying must be with a volatile solution such as freshly

prepared ammonium acetate or STEM water. Because ammonium acetate is a

poor buffer, it can be used at any desired pH.

2.2. Grids and Films

1. We use titanium grids for the STEM. This is for ease of handling in the

microscope and for their good thermal behavior. The supporting film (step

3) is thin carbon. The grids are at liquid-nitrogen temperature in the micro-

scope to prevent contamination and reduce radiation damage to the speci-

mens. Carbon and titanium have similar thermal properties, so there is no

differential contraction or expansion that would result in tearing of the film.

Also, compared to copper grids, titanium grids are flat and chemically inert.

However, to mimic STEM conditions in a conventional TEM, there are

many choices for grids.

2. The slots in our titanium grids are too large to support the thin carbon film neces-

sary for STEM samples, so we put a “holey” film over the slots. This is a plastic

film made on glass slides under controlled humidity conditions so little droplets

of condensation determine the hole size, which is approx 5–10 mm in diameter.

The plastic holey film is heavily shadowed with carbon for strength. Again, to

mimic our conditions, there are many available choices of grids with support for

small areas of thin film.

3. The thin carbon film is critical to simulating STEM conditions for samples.

Because carbon film behaves like activated charcoal in adsorbing everything from

the atmosphere, it is made so that the side to which the sample is adsorbed has

never seen air. We make it by lightly shadowing carbon (for the STEM, it needs

to be 2–3 nm thick) onto a piece of freshly cleaved rock salt in an ion-pumped

vacuum system. (See Note 10.) We buy NaCl crystals from Bicron. We cleave

them with a Weck safety razor blade, which is placed on a corner edge of the

crystal (to give a piece several millimeters thick) and tapped with a hammer. The

freshly cleaved side is lightly breathed on before putting it face down in the bell

jar. The carbon rods (from Fullam) are located below the rock salt in the shadower

so that cinders do not drop on it. (See Note 11.)

4. We often use poly-lysine pretreated grids for complexes on dsDNA, which does

not stick well to carbon grids. However, if the grids are pretreated with poly-

lysine, dsDNA binds very well. We use poly-lysine from Sigma, approx 3000

Daltons, and dilute and freeze away aliquots at 10 µg/mL.

5. We sometimes fix complexes with low concentrations (approx 0.1%) of glu-

taraldehyde for relatively short times (approx 15 min) before applying them

to the grid. The sample washings will stop the fixation so that it does not

have to be quenched. (See Note 4.) We start with a new ampoule of 8% EM

grade from Polysciences. We usually open a new ampoule, but if the air is

flushed out with dry nitrogen gas and it is resealed, it can be kept for a few

days for further use.

STEM of DNA–Protein Complexes 593

2.3. Concentrations of Complexes and Components

1. Only small quantities of sample are needed for examination in the STEM, but it

is useful if the concentration is relatively high. Because different samples absorb

to the thin carbon film differently, a concentration series for a new specimen is

made. Twenty to 50 µL of a sample at 100 or 200 µg/mL is adequate. (See Note 12.)

2. It is often a good idea to look at the components tht go into a complex separately.

On poly-lysine pretreated grids, a dsDNA concentration of 1–5 µg/mL is plenty.

Depending on its size, a protein concentration of 50–100 µg/mL is good. Again,

only 10–50 µL is needed.

3. The ideal sample is frozen in aliquots to be thawed and used as needed. However,

many samples cannot be frozen. They can be shipped overnight on wet ice and

the grids can be made on the day of arrival (if it has been arranged). The grids are

stored under liquid nitrogen until microscope time is available.

4. Occasionally, a complex is unstable, either over time or shipping conditions.

Sometimes, frozen aliquots of the components can be shipped, but the complex

itself cannot be frozen or it may not be stable overnight. Under those circum-

stances, we can assemble (and incubate) a complex just prior to putting it on the

grids. We need a detailed protocol and all of the components.

3. Method

3.1. Specimen Preparation

Details of specimen preparation for grids for the STEM are given in this

section. Specimens for the STEM are prepared by the STEM staff (see Note 5),

but knowing the details enables users to understand the potential problems.

A user with EM experience in other ways of visualizing proteins or DNA or

complexes can mimic these conditions to screen their samples in a conven-

tional TEM.

3.1.1. Wet Film, Hanging Drop Method

1. The thin (2–3 nm) carbon film, which has been shadowed onto freshly cleaved

rock salt, is floated off the crystal onto a dish of STEM water (see Note 6).

2. The grids, covered with holey film (and for the STEM, in rings and caps) are

placed face down on it for approx 1 h.

3. A grid is picked up from above and retains a droplet of water. It is washed and

wicked with STEM water two or three times, never being allowed to dry. Water

is applied from one side with an Eppendorf pipet and wicked from the other side

with filter paper.

4. Three microliters of tobacco mosaic virus (TMV) at 100 µg/mL is injected into

the drop and allowed to adsorb to the thin carbon film for 1 min.

5. The grid is then washed and wicked two or three times with injection buffer for

the sample. Three microliters of the specimen is injected into the drop and allowed

to adsorb for 1 min.

594 Wall and Simon

6. The grid is washed with sample buffer a couple of times, followed by washes

(approx 10), ending with washes of 20–50 mM ammonium acetate (or, rarely,

water). The intermediate washes depend on the sample buffer (e.g., high-salt buff-

ers require washes with high ammonium acetate). (See Notes 8 and 9.).

7. For freeze-drying the grid, after the final wash the liquid is briefly blotted between

two pieces of filter paper to a thickness of a few micrometers and rapidly plunged

into liquid-nitrogen slush to freeze it.

3.1.2. Poly-lysine-Pretreated Grids

The dsDNA does not bind well to carbon films. However, if the grids are

pretreated with poly-lysine, dsDNA binds very well to them. Poly-lysine

grids that have been made by this method contribute no additional detect-

able background.

1. The thin carbon film is picked up as above and washed with water.

2. Three microliters of poly-lysine at 10 µg/mL is injected into the drop and allowed

to adsorb to the carbon film for 1 min.

3. The grid is then washed with water, approx eight times, and allowed to air-dry.

4. Before use, 3 µL of water is put on its surface before the TMV is injected. Because

TMV also binds very well, a solution of 10 µg/mL should be used for these grids.

(See Note 13.)

3.1.3. Fixation

In any multicomponent system, one part may bind more strongly to the car-

bon film substrate. The rest of the complex may be coming apart through

Brownian motion as described earlier. Frequently, some kind of crosslinking

or fixation is required to keep the complex together.

The best results are when the users have been able to fix or crosslink their

sample before sending it to us. Sometimes they have determined the minimal

fixation or crosslinking conditions for their sample, but it needs to be done just

prior to applying the sample to a grid. Unless something very unusual is called

for, we can do that.

If the complex is falling apart and optimal fixation conditions to preserve it

are not known, we will often try a brief glutaraldehyde fixation on the sample

just prior to applying it to the grid. The washes involved in specimen prepara-

tion will stop the fixation without having to worry about quenching conditions.

Sometimes, this will help preserve the complexes, but it may not.

3.2. STEM Operation

The STEM is usually run by a trained operator to assure optimum data qual-

ity and efficient use of microscope time. Frequent users can obtain training on

request, but the alacrity of most STEM experiments makes this unnecessary.