Radiation damage limits the usefulness of TEM data regardless of the initial quality of the specimen. Radiation damage rather than the instrumental resolving power is the main limiting factor in obtaining high resolution images of biological molecules.
Most biological specimens tolerate an exposure of no more than 100-500 e-/nm2. The highest resolution features of the specimen are already affected at electron exposures as low as 10 e-/nm2 or less (Table 1). The structural effects observed include:
- mass loss as great as 50% Damage is proportional to the TOTAL DOSE, which is the product of the beam current density (dose rate) and the exposure time. For most specimens, damage is not dose-rate dependent. Dose is a measure of the energy absorbed by the specimen. In radiology, the conventional unit is the rad, which is 100 ergs absorbed/gm. What is generally measured in TEM is the electron flux in coulombs/cm2 or e-/nm2 (recall that 1e- = 1.6x10-19 coulomb). Multiplication of the flux by the beam voltage gives the incident energy. However, using this quantity as a measure of the dose can be misleading, because with a thin specimen and a fixed electron flux, the absorbed energy falls as the beam voltage rises. The conversion to rads is therefore not straightforward as it depends on the energy loss per unit length of passage through the specimen, hence on specimen atomic weight and beam voltage. For carbon and carbonaceous materials a rough conversion from electron beam flux to radiation dose (energy) is given by 1 coulomb/m2 (or ~6 e-/nm2) at 100 kV = 4x107 rads. 109 rads is sufficient to destroy the original properties of most organic materials, while a dose of 106 rads will kill the most resistant of all living things and cause the biological inactivation of most enzymes. The electron dose leading to complete molecular disorder in an unstained specimen is ~100 e-/nm2 at 80 kV and ~250 e-/nm2 at 500 kV. The minimum current density in order to barely visualize the object at the fluorescent screen is ~1 e-/µm2/sec at the screen, which is a dose rate at the specimen of 400 e-/nm2/sec if the magnification is 20,000X. Thus, there is not even enough time to visualize the specimen let alone focus and record an image. In a two second exposure (which would produce an optical density of about 2.0 on a photographic emulsion with a film speed of 1.0 µm2/e-), the amount of energy at the specimen would be about 5x109 rads which is approximately equivalent to the energy delivered from a 10 megaton Hydrogen-bomb exploding 30 meters away! II.B.3. Primary Effects of Radiation Damage to Biological Specimens a. Introduction Specimen damage is primarily caused by ionization resulting from the inelastic interaction of electrons with the orbital electrons of the organic material. This, in turn, leads to rearrangement of chemical bonds, formation of free-radicals and diffusion of the fragments. Whereas elastic interactions produce image contrast but no damage, inelastic interactions can produce permanent changes. The two most important structural changes that occur in biological molecules include: - molecular fragmentation leading to mass loss as the material liquifies or vaporizes Table 1. Imaging requirements and radiation damage at 1MeV. (From Glaeser [1975]12:215) b. Interactions between the electron beam and the specimen Radiation interactions that result in bond rupture are the primary cause of damage to biomolecular structure. Specimen heating does not occur except at much higher electron intensities than normally used. Even at higher intensities heating would only occur subsequent to an extensive amount of direct damage. The three primary interactions between the electron beam and the specimen are excitation, ionization, and displacement. These are all virtually temperature independent and occur very rapidly, on the order of 10-14 seconds. An excited molecule may dissipate its energy by several different routes (secondary effects), not all of which are destructive. The energy of the excited molecule can be dissipated in many ways: - Electron, x-ray or light emission: e.g. electromagnetic radiation as fluorescence or phosphorescence in the x-ray, ultraviolet or visible regions When electrons in the beam knock off atomic orbital electrons in the specimen, they cause ionization and the resulting forces may cause changes in the specimen by inducing bond linkages or chemical chain scission. Bond rupture and related molecular damage associated with excitation or ionization is normally the limiting consideration in radiation damage. The cross section for atomic displacement is very much smaller than that for ionization and excitation so it will only be expected to play a significant role in organic materials that are very resistant to damage following ionization. c. Voltage dependance The mechanism and amount of energy transfer from the beam to the specimen depends primarily on the voltage. The inelastically scattered electrons, which interact with the orbital electrons of the specimen atoms and transfer large amounts of energy, are considered to be primarily responsible for the energy transfer that occurs from the beam. If the composition of the sample and the current density of the beam are kept constant, the beam-to-specimen energy transfer can be minimized by increasing the voltage and decreasing the sample thickness. At 100 kV it has been calculated that an electron passing through an organic specimen 0.1 µm thick sustains, on average, one inelastic collision (of about 32 eV; Note: the energy of a C-C single bond is about 10 eV; 1 eV = 1.6 x 10-19 joule). d. Steps in the process of beam induced damage The three major steps in the process of beam damage include: - Physical step: This lasts for less than 10-13 sec and involves the inelastic scattering of an electron and results either in direct ionization or excitation of an atom. This step is virtually independent of temperature. e. Resistance to beam damage The molecular parameter which determines a molecule's ability to resist damage by ionizing radiation is the resonance energy per p electron. The larger this quantity, the more delocalized are the p electrons and the easier it becomes for the excitation energy to be dissipated within this p electron pool without bond damage. Molecules composed of extensively conjugated, p electron systems exhibit a greater thermal stability and tend also to be less damaged by ionizing radiation. Aromatic hydrocarbons exhibit a high resistance to radiation and aromatic compounds have a protective effect on aliphatic compounds. II.B.4. Secondary Effects of Radiation Damage a. Nature of the damaged, beam-resistant product Alterations that occur in the specimen take place very rapidly, so that after a few seconds the specimen acquires complete morphological stability in the beam. One may conclude that, since the image does not change with time, that after a given dose, the specimen approaches a steady state with the beam, at least on a morphological level and perhaps also on a chemical level. Three experimental data corroborate this: - Mass loss reaches a plateau despite continued irradiation. Electron irradiation alters or destroys secondary and tertiary structure in molecules, causes loss of specific groups, increases unsaturation, causes cross-linking and/or scission of large molecules, and alters elementary composition. Cross linking increases the molecular weight while scission decreases the molecular weight and produces volatile fragments. Which of these processes dominates depends on the structure of the compounds. The final product is determined to a certain extent by the molecular characteristics of the initial product and its molecular environment. b. Secondary effects Electron bombardment results in one or more of the following effects: - Chemical and physical changes The amount of energy transferred from the beam electrons to the specimen electrons is so far in excess of chemical binding energies that no substance can be regarded as completely stable under a high voltage electron beam. Stability or instability depends on whether, after excitation of electronic energy levels, the substance falls back to its original structure or to some new configuration. Most organic materials are probably altered so as to be chemically quite different, even though the structural features within TEM resolution may not be obviously changed. The rate of chemical change can be much reduced by lowering the beam intensity to the minimum required for observation at low magnification. It sometimes helps to monitor beam induced structural changes by recording several images at time intervals. Ionizing radiation causes a variety of molecular changes. C-H bonds are very sensitive to radiation while C-C bonds are more resistant. The number of C=C bonds increases with increasing irradiation dose owing to the rupture of C-H bonds and of side chains. The loss of side groups and preservation of the carbon backbone means that most organic molecules show a higher carbon content as radiation exposure is increased. The final product is predominantly, but not pure, carbon. On losing hydrogen, the organic molecule acquires more double and triple bonds. This leads to a change of bond length and angle, and thus conformational change. Amino acids appear to be most sensitive at the a-carbon where a free-radical forms. Large molecules respond to irradiation with cross-link formation or scission. The ion or radical produced on a large molecule may react to form a very stable covalent bond with another radical or ion on an adjacent molecule. Such cross-linking leads to the production of large and tightly bound molecular aggregates. All organic molecules appear to undergo mass loss upon irradiation in a vacuum. The mass loss is dependent on: - the composition of the specimen Mass loss is believed to result from the fracture or scission of the specimen molecules. Since continued scission alone would progressively reduce the mass of the specimen in the TEM, the stable product we see is either resistant to the beam or it is undergoing or has undergone cross-linking. In the TEM therefore, although specimen scission definitely occurs as seen in mass loss, the predominant reaction must be cross-linking. The rate of mass loss is initially rapid but, with continued irradiation, levels off and reaches a plateau. The variation of loss of mass with beam intensity is attributed to heat since it is believed that, if molecular fragments have a greater mobility, there is a greater opportunity for loss of substance to the vacuum. Mass loss generally occurs in the first few seconds of illumination. As the energy of the beam electron is reduced as it passes through and interacts inelastically with the specimen, a point is reached at which it can no longer electrically excite. At that point it imparts energy to lower energy states: vibrational, rotational, and translational states (i.e. heat). The temperature of a specimen is a function of the difference between heat input and heat dissipation. The parameters of importance are: - rate of energy delivery (current density and voltage) As more energy is put into the specimen per unit time, heat becomes an increasingly important factor. This occurs when the current is high, the voltage is low, and the specimen is thick and contains elements of high atomic number. Calculations indicate that the specimen temperature can reach as high as 300°C. Cooling of the specimen can be effective only if it is supported so as to ensure good thermal conductivity. The most effective way to reduce heating is to keep the current density to the minimal practical value for observation on the final screen. It is possible that higher mass loss in heated specimens is due to greater mobility of the molecular fragments. Fragmented monomers could be cross-linked in cold specimens before leaving the object, whereas in heated specimens they would evaporate before cross-linkage could occur. The mechanism of the effect is believed to be an uneven flow of charge onto and from the sample. During irradiation the electrons and ions produced can lead to a radiation-induced current proportional to the radiation intensity. Charge effects probably accelerate sublimation through the ejection of small fragments as a result of electrostatic repulsion. Charge effects result mainly in the movement of particles with respect to one another or deterioration or movement of the image. The frequent practice of applying a thin layer of evaporated carbon to one or both sides of the specimen (carbon sandwich) may produce its beneficial effect by providing increased conductivity and thereby eliminating the charge. Specimens exposed to intense electron beams in a vacuum generally become coated with decomposition products of the residual vapors of oil and grease in the vacuum chamber. The principal vapors are hydrocarbons which break down under electron bombardment, deposit on the area illuminated, and cross-link to form polymers. If these vapors are trapped away from the specimen using an anticontaminator, water vapor becomes the chief offender. Water molecules become ionized in contact with the specimen, and carbon specimens become oxidized and etched away. The chief source of water is the photographic emulsion. Many crystalline specimens are more resistant to irradiation damage than amorphous specimens. Crystalline destruction is ascribed primarily to ionization and thermal effects. Protein crystals may contain a considerable amount of water of crystallization which does not leave the crystals in vacuum at room temperature. If, however, such crystals are heated in a vacuum due to beam effects, the water of crystallization is driven out and the crystals shrink or become distorted. II.B.5. Ways to Measure Damage/Critical Dose A variety of criteria have been used to measure radiation damage, including total mass loss, loss of specific elements, loss of crystalline structure, and changes in the infrared, visible, or ultraviolet spectra (including energy loss spectra). Most often, the extent of damage is assessed using radiation-induced changes in electron energy loss spectra and in electron diffraction patterns. As measured by infrared spectroscopy, in the dose range 60-200 e-/nm2, the secondary structure of proteins is completely randomized. a. Electron diffraction vs. energy loss spectroscopy Whereas electron diffraction records primarily the loss of order in a periodic specimen, energy loss spectrometry shows the rate of destruction of molecular organization. Decay during irradiation usually proceeds much more slowly than the fading of the diffraction pattern, indicating that bonds between molecules are in general broken more readily than those within a molecule. Low-lying (<50 eV loss) energy-loss spectra are determined by the p and s electron energy levels of the molecule. Thus, changes in the spectra as a function of the incident electron dose can provide information about molecular structural damage (Fig. II.79). Changes in the diffraction patterns provide information about crystal structure (or weak binding force) damage (Fig. II.80). b. Critical dose The critical dose is defined as that dose at which the intensity of a given peak in the electron diffraction pattern falls to 1/exp (i.e. 37%) of its original value at zero dose (Fig. II.81). For unstained protein crystals at room temperature this is about 50-150 e-/nm2. The critical exposure for organic molecules predominantly composed of saturated bonds at 100 kV ranges between 50-500 e-/nm2. The improvement in critical exposure at 1 MeV is only about a factor of 2-3 over that found for the same material at 100 kV. Model calculations indicate that the percent reduction of intensity in diffraction patterns is roughly equal to the percent damage developed in the specimen. The critical dose for molecular damage, as indicated by energy-loss spectra, is about ten times greater than the critical dose for crystal structure damage as indicated by the fading of the diffraction pattern. For l-valine the dose for complete fading of the electron diffraction pattern is 80 e-/nm2 at 80 kV and 200 e-/nm2 at 500 kV. This corresponds to a dose for complete disorder of about 30 eV or one average ionizing event per unit cell. Since there are 4 molecules/unit cell, on average, ionization of one molecule per cell is sufficient to disorder the crystalline structure. For adenosine the dose is 90 eV/molecule or three ionizing events/molecule. The greater resistance of adenosine might partially be explained by its having a conjugated p-electron system. Because of the delocalization of the outer shell valence electrons, structures of this type are relatively stable to single electron ionizations. The radiation damage of stained catalase is of a fundamentally different nature than for unstained valine or adenosine. The dose corresponds to about 200 ionizing events/nm2 or 6000 per molecule. A significant portion of the matter within the unit cell of the structure changes to some more stable configuration as a result of the irradiation. This would account for the changes in the relative intensities of the diffraction maxima that remain after irradiation. The distribution of matter becomes more disordered as well, since only the lower-resolution diffraction reflections remain. This redistribution of matter probably includes a significant redistribution or aggregation of stain molecules. c. Dose rate Quantitative measurements of radiation damage in crystalline specimens of l-valine, adenosine, and uranyl-stained catalase, using fading of electron diffraction patterns, indicates there is no dose rate effect, such as specimen heating, in the range 0.5-50 e-/nm2/sec current density at the specimen. d. Mass loss Electron autoradiography has been used to measure the amount of mass loss in biological specimens irradiated with electrons. Loss of 14C due to electron irradiation has been measured, for example, in labeled T4 bacteriophages, ribosomes, R17 virions and, E. coli. The specific activity of labeled specimens was measured on specimens either exposed or unexposed to the beam. It was found that macromolecular biological samples are severely damaged during standard operations: from 15-40% of the mass of the sample may be lost in a 30 second exposure to the beam. During irradiation the remaining material became less and less sensitive to further carbon loss. No difference was found for positively-stained verses unstained samples. A small decrease in the damage rate occurs when the beam spot size is reduced, probably leading to less specimen heating. It was found that surface migration of molecular fragments and adsorbed molecules is involved in the process of beam damage. At liquid helium temperature there was no perceptible carbon loss with irradiation. The rate and extent of mass loss with negatively-stained or metal-shadowed specimens has not been reported. II.B.6. Procedures to Reduce Radiation Damage There are several ways to help minimize the radiation damage to biological specimens. Some of the more commonly used methods include: using fewer electrons, lower temperature, higher voltage, reducing contamination, coating specimens with a thin layer of carbon and using carbon support films, and preirradiating the specimen. a. Minimize the number of electrons The most obvious method to reduce radiation damage is to use fewer electrons either by cutting down on the exposure time or by reducing the beam intensity. A basic rule is to use the minimum magnification that will reveal detail of a given size (determined by the resolution of the photographic emulsion), thereby enabling the shortest possible exposure time to record the details with sufficient contrast on the photographic emulsion. A number of minimal exposure techniques have been developed for recording images of biological specimens with reduced beam damage. These techniques involve focusing on one region of the specimen but photographing another. In this way, exposure of the area of interest on the specimen to the electron beam only starts when the photograph is being taken. Initial minimal exposure techniques required the TEM grid to be translated a few microns so focusing could be performed away from the specimen. However, the translation stage movements do not remain perfectly planar to the required accuracy, hence the position of focus can change when the specimen is then translated back to the optic axis. Recent techniques allow focusing to be performed with the specimen in position so no further stage movements are necessary. In a method that was widely used in the 1970's, a second, small viewing screen positioned off axis, was used for off-axis focusing (away from the specimen). Then, with the photographic plate already in position, the beam was spread out to cover the whole area of the screen and the photograph was immediately recorded. Modern microscopes include additional beam deflection circuitry for performing the minimal irradiation procedure. In the Philips EM420, for example, the beam can be accurately aligned so it can be deflected to a region a few microns away from the specimen. After passing through the region off the specimen, the beam is redeflected back to the optic axis so the image of the area used for focusing appears in the center of the viewing screen. When focusing and astigmatism corrections are made to satisfaction, and the exposure button is pushed, the beam is automatically recentered and expanded to a predetermined amount and an image of the desired area is recorded. After the exposure is finished, the beam is again automatically redeflected away from the specimen to prevent further irradiation. Since photographic emulsions are nearly perfect recording media, image intensifiers are of little value in preventing radiation damage. They can, however, aid focusing at low illumination levels. (See Sec. I.E.8). b. Low temperature Chemical bonds are broken by electron impact regardless of the temperature At low temperature some of the molecular fragments remain fixed in position and the image retains the topography of the original object after more irradiation than at room temperature, but when the specimen warms up, damage is no different than if the irradiation had occurred at room temperature. Some evidence suggests that, for a given dose, damage increases with increasing flux. The trapping of highly reactive radicals at low temperature might also serve to reduce structural damage that would otherwise occur as a secondary effect due to chemical reactivity. For small biological molecules such as crystalline l-valine and adenosine, examined at 190 kV and at 15-20°K, there was no significant effect, within experimental accuracy, of low temperature preserving the crystalline structure as measured by the disappearance of the electron diffraction pattern corresponding to the loss of crystalline order. There was also no dose rate effect with these specimens. The results seemed to imply that high concentrations of trapped radicals at low temperature might lead to large reactivity between adjacent groups. Also, the loss of a substantial percentage of the hydrogen atoms, which apparently does still occur even at 4°K, might cause a structure to "relax" to a new and possibly more disordered configuration. Radiation-induced scission of chemical bonds may make it possible for the fragments to relax into previously inaccessible minima of the intermolecular potential energy surface with concomitant disorder. Thin crystals of the crotoxin complex (a macromolecule) embedded in glucose are preserved with a mean electron exposure about 10-fold higher at 4°K than at 300°K. These results are explained by the existence of "caging" effects after the specimen undergoes primary damage by ionization and excitation. "Caging" reduces the rate of migration and diffusion of the damaged fragments and also minimizes the occurrences of chemical reactions among the newly generated free radicals and ions. The electron exposure required to completely destroy the electron diffraction pattern of frozen-hydrated catalase was found to be more than ten times greater than the corresponding critical exposure found for wet hydrated catalase at room temperature. At room temperature catalase crystals lose the 6.9 nm lattice reflection in diffraction patterns after 500 e-/nm2. At very low temperature, the same reflection is still present after 450,000 e-/nm2. The lifetime of glucose-embedded catalase crystals is hardly distinguishable from that of frozen-hydrated catalase when both are observed at liquid nitrogen temperature. Radiation damage protection found for small molecule organic compounds is generally rather limited or even non-existent at low temperatures. However, hydrated macromolecular materials show as much as a 10-fold reduction at low temperature in the rate at which radiation damage occurs, relative to the damage rate at room temperature. A major consequence of examining specimens at low temperature is that mass loss is greatly reduced. A dose of 1000 e-/nm2 can result in a mass loss of 20-80% at room temperature, whereas at very low temperature, specimens are more stable and mass loss is suppressed. Sometimes a latent dose effect is observed in which no major change in electron diffraction occurs during the first part of the radiation exposure. This effect appears to be related to a trapping of the primary products of radiolysis which eventually lead to destructive chemical reactions. c. Reduce contamination Reduction of contamination by use of anticontaminator blades in the vicinity of the specimen helps to reduce radiation damage by removing molecules which can become ionized in the beam and subsequently attack the specimen. Hydrocarbons from the diffusion pump oil and greased gaskets generally cross-link to the specimen under the influence of the beam, thus adding mass non-specifically to the specimen and reducing its contrast. Water molecules from insufficiently desiccated photographic films, become ionized in the beam and etch the specimen (i.e. reduce its mass). A suggestion has been made that specimens be placed in the microscope with the support film side facing down the column and the specimen facing the electron source to help reduce etching effects (i.e. mass loss) since the support film would act to shield the specimen from water molecules entering the column from the camera chamber below. d. Carbon stabilization of specimens and carbonsupport films Coating a specimen with a thin layer of evaporated carbon has a pronounced stabilizing effect, reducing the rate of mass loss and prolonging the lifetime of diffraction patterns. Studies have shown that crystals mounted on plastic films are more beam sensitive than crystals mounted on carbon films. This presumably reflects the more efficient heat and electrical conductivity of carbon compared to plastic films. For the same flux, damage occurs more rapidly when a large surface of the specimen is irradiated. For thin specimens mounted on carbon films, beam damage is independent of the flux in the range 1-300 e-/nm2/sec. Under these conditions, specimen heating is probably negligible. e. High voltage The use of higher operating voltages to reduce radiation damage has been a somewhat controversial matter. While it is true that specimen lifetimes at high voltages (1 MeV) are about 2.5 times greater than at conventional voltages (50-100 KeV) because faster electrons deposit less energy (i.e. fewer inelastic scattering events) in a specimen of given thickness, this is offset by the reduced amplitude contrast at higher voltage due to fewer scattering events (elastic and inelastic) in the specimen and weaker interaction with the photographic emulsion (fewer developed silver grains, therefore less optical density and lower photographic contrast). Total inelastic scatter from carbonaceous material exceeds the elastic scatter by about 3:1. The ratio is unaffected by acceleration voltage and no further reduction in beam damage is expected from the use of high voltage (1 MeV) for biological electron microscopy. f. Preirradiation of specimens Several studies have shown that a small dose of electrons on some specimens (especially thin sections) helps stabilize them against further effects of irradiation. This procedure presumably works by initially favoring cross-linking reactions at low beam intensities, thus enabling specimens to withstand higher doses without mass loss due to bond scission events. Preirradiation with a small dose (<100 e II.B.7. Relation between Contrast, Resolution and Radiation Damage As the electron exposure is reduced, statistical fluctuations from one picture element to another can be much greater than the inherent change in image intensity from one picture element to another, i.e. the statistical fluctuations may greatly exceed the inherent contrast. Thus, by reducing the electron exposure, the situation may be such that the specimen detail is well preserved, but it cannot be observed because of the noisy quality of the image. At high resolution there does not seem to be any way to escape this dilemma except through the use of image averaging techniques, which are especially powerful when applied in the study of crystalline specimens (Sec. III, taught in BIO695 Spring 1995). A fundamental limit to resolution exists because of poor statistics in the measurement. Image features with low contrast are difficult to see at low electron fluxes because the statistical fluctuation of image intensities may exceed the small variation of intensity associated with the intrinsic contrast. Since the incidence of electrons at the image plane is a random process, then the total number of electrons passing into a given image point (picture element) is n and the statistical fluctuation, or counting error is ÷n. For a low-contrast image feature to be resolved, the inherent contrast must exceed the statistical fluctuations by the minimum acceptable S/N ratio. Radiation damage can destroy the object to such an extent that an image with statistically well-determined contrast actually may no longer contain meaningful information about the object at the desired resolution. II.B.8. Radiation Effects in Negatively-Stained Specimens The effects of the electron beam on negatively-stained biological specimens has been studied by Unwin (1974) for samples at room temperature and Glaeser (1975) at low temperatures (see Radiation Effects reading list). a. Negatively-Stained samples at room temperature Unwin analyzed the same negatively-stained TMV stacked disk specimen subjected to successively higher electron doses and showed that the stain molecules migrate and redistribute themselves over the virus particle surfaces during irradiation (Figs. II.82 - II.84). The morphology of the stain accessible regions of the specimen strongly influences the path of stain migration. With exposure, uranyl ions are reduced by incident electrons from their initial 6+ oxidation state to the 4+ state. The stain shrinks quite drastically while it is irradiated (Figs. II.83 and II.84). The density of uranium dioxide crystals produced during irradiation is 11.0 g/cm3, which, when compared to the density of the acetate before irradiation (2.9g/cm3), shows there is a considerable change in the density of the "stain" upon irradiation. It appears that the stain, before it becomes completely microcrystalline, behaves in an almost liquid-like manner. The protein appears to be morphologically quite well preserved and stable. Migration of stain is not accompanied by a significant loss of resolution, thus it may involve an orderly redistribution. The reason for the apparent stability of the protein seems to be that the major alterations taking place in it occur very rapidly, at doses smaller than those required for a high resolution micrograph even under low dose conditions (²1000 e-/nm2), and that when the protein is being observed it has more or less reached a steady state condition, dissipating further incoming energy non-destructively. The tertiary structure of "unprotected" protein, which is strongly dependent on hydrogen bonding and other weak interactions, is normally very rapidly disrupted by the radiation or by associated charging effects. The stain may prevent this disruption in the early stages by constraining the polypeptide chains close to their original positions and enabling extensive cross-linking to take place between them. A stable, highly cross-linked, structure would then be formed, the gross morphology of which is unlikely to be identical to that of the original protein, but at about 1 nm resolution or so should bear a close resemblance to it (but see Fig. II.85). There are two distinct stages of irradiation: - An initial fixation stage (<500 e-/nm2) involving considerable atomic rearrangement and associated structural alterations in both stain and protein. b. Negatively-Stained samples at low temperature The fact that stained specimens are more sensitive to radiation damage at low temperature than at room temperature suggests that the mechanism of damage is somehow different from that which occurs in unstained specimens, since the latter tend to show a modest decrease in sensitivity at low temperature. It may be that the chemical radicals and low molecular weight species become trapped in the specimen until they reach very high concentration, at which point they come in contact with one another and a relatively rapid reaction may take place. The interaction of these products of radiolysis at high concentration may have a more cataclysmic effect upon the structure than if the products could have been consumed or released at the same rate that they were produced. Presumably the latter is what occurs at room temperature. II.B.9. Radiation Effects in Frozen-Hydrated Specimens This was discussed in Sec. II.A.5.c. 
Ions or radicals result when enough energy is absorbed by an orbital electron to become free or reach a higher energy (non-bonding) orbital. Subsequently that molecule has either a charge (ion) or an unpaired electron (radical). The fates of these species are known as the secondary events of radiation, and it is these reactions, not the primary events, that cause material damage.
Fig. II.79. Infrared absorption spectra of non-irradiated (top) and irradiated (bottom) polyamide. (From Bahr, Johnson, and Zeitler [1965] Lab. Invest. 14:377) 
Fig. II.80. Changes in the electron diffraction pattern of frozen-hydrated catalase crystals resulting from radiation damage. (a) The initial diffraction pattern extends to 2.8 Å. The ring present in the pattern at a spacing of 3.67 Å is due to ice condensed on the surface of the specimen from water vapor present in the column of the microscope. (b) The pattern recorded after an exposure of 2.5 e-/Å2. (c) The pattern recorded after an exposure of 5.0 e-/Å2. (d) The pattern recorded after an exposure of 11 e-/Å2; the pattern still extends to 8.5 Å resolution. (From Taylor and Glaeser [1976] J. Ultrastruc. Res. 55:448) 
Fig. II.81. The intensities (on a logarithmic scale) of some typical reflections in (a) catalase and (b) the purple membrane electron diffraction pattern, plotted as a function of electron dose. The dotted curve is for one of the few reflections which actually increases in intensity during the intitial exposure period. (Such reflections are at variance with the more typical behavior, over the dose range shown, of exponential decay.) The data for (b) were taken from the "ring" diffraction patterns produced by a large number of native membranes several layers thick, so that the intensity of the (4,3) ring is contributed by both the stronger (4,3) and the weaker (3,4) lattice reflection, and so on. (From Unwin and Henderson [1975]J. Mol. Biol. 94:425) 
Fig. II.82. A schematic representation of the behavior of uranyl acetate and PTA under the influence of electron bombardment. Both uranyl acetate and PTA scattering centers break up under the electron beam, but the uranyl acetate shows more cohesion. Uranyl acetate is also transformed into small granular lumps. It is assumed that the support film (thick line) is not changed under the electron beam. (a) Uranyl acetate or PTA stain profile before irradiation; (b) PTA stain after electron bombardment; (c) Uranyl acetate after electron bombardment. (From Hayat & Miller, p.16)
Fig. II.83. Suggested influence of an asymmetrical groove in the protein on the radiation changes produced in the stain near the outer surface. (a) Initial conditions, the stain near the outer surface (i.e. along the broken line) being centered to the right of the base of the groove; (b) the effect of contraction of the stain about its center of mass; (c) effect of surface energy included. The contraction produces a small net displacement of the stain towards the left near the outer surface, and this is greatly enlarged by the surface energy effect, causing the stain eventually to be centered directly above the base of the groove. (From Unwin [1974] J. Mol. Biol. 87:657) 
Fig. II.84. Radiation induced changes induced in thin films of uranyl formate stain. Image doses: (d) 16 e-/Å2. (e) 625 e-/Å2. The images are at the same magnification and the markers on the stain film are tungsten oxide particles, which have come closer together in (e) as a result of its radiation induced contraction. (From Unwin [1974] J. Mol. Biol. 87:657)
Stained catalase, examined at low temperature, has a curiously higher radiation sensitivity than when studied at room temperature. The electron diffraction pattern of uranyl-stained catalase at room temperature fades down to a stable pattern at a resolution of about 2.5 nm. At liquid helium temperature, however, the diffraction pattern fades down to a resolution of about 7 nm and, in some cases, all of the diffraction spots are lost. 
Fig. II.85. Images pairs of uranyl acetate-stained TMV, recorded with minimal irraditaion methods at about 20e-/Å2 and 200 e-/Å2, respectively. The first image in each pair clearly shows the 23 Å helical repeat of the protein subunits in the TMV rods. Finer details can also be seen in the lower-dose image of the TMV stacked disc aggregate (left edge of top and bottom right image pair). Magnification bar = 500 Å. 
Fig. II.86. Dose series of frozen-hydrated SV40 virus showing how bubbles form first over the carbon film and then in the virus particles over the hole. The water over the hole forms bubbles after continued irradiation. (From Baker et al., [1985] Proc. Elec. Microsc. Soc. Amer. 43:316)