- Email: [email protected]
A technique for the electron microscopy of protein-free particle suspensions by the negative staining method
© 1970 by Academic Press, Inc.
J. ULTRASTRUCXURERESEARCH32, 107-117 (1970)
107
A Technique for the Electron Microscopy of Protein-Free Particle Suspensions by the Negative Staining Method I MAGDALENA REISS1G AND NTANLEYA. ORRELL
Department of Pathobiology, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, Maryland 21205 Received August 21, 1969, and in revised form January 5, 1970 A method has been developed for the electron microscopic study of proteinfree particle suspensions by the negative staining method without the addition of substances that would lower the surface tension of the preparation to facilitate spreading. It involves treating the carbon supporting films with a high voltage glow discharge under carefully controlled conditions, before spraying on them the suspensions to be observed. Its applications include particle counts in microdrops with the addition of latex standards and evaluation of particle size distributions. It is particularly useful in those cases when one wishes to avoid the use of additives, such as detergents or soluble proteins, which may react with the particles under study or with the negative staining solution. Use of the negative staining method (2) for the study of particulate materials by electron microscopy, requires that the preparation be spread on the supporting film in a thin layer (9). Since the aqueous solutions of phosphotungstate and other negative stains generally used do not wet adequately the supporting films, the method has been successful only when the surface tension of the particle suspension was low enough to permit adequate spreading. Most biological preparations contain some soluble protein, either as the substance under study or as a contaminating impurity, which act as a surfactant. If such an agent is not present, either albumin (10) or detergents (15) can be added for this purpose. Another approach to this problem would be to devise a method to modify the surface of the supporting film in such a way that it would be readily wettable by the preparation, even if its surface tension was high. The present report is concerned with the description of such a technique, which was devised for the study of protein-free suspensions of glycogen. It is based on the principle, known for some time, that glass and metal surfaces can be cleaned and made surface active by exposing them to a high voltage glow discharge passed between metal electrodes in a continuously exhausted vessel (4). i This investigation was supported in part by the United States-Japan Cooperative Medical Science Program administered by the National Institute of Allergy and Infectious Diseases of The National Institutes of Health, Department of Health, Education, and Welfare.
108
REISSIG AND ORRELL
MATERIALS A N D METHODS The following procedure was developed for preparing wettable carbon films: pure carbon films evaporated on freshly cleaved mica by the method of Hall (3) are mounted on copper grids and fastened with tape to glass slides. They are then exposed to a high voltage glow discharge in a standard vacuum evaporator with a modified bell jar and a power supply for the glow discharge. The power supply for the discharge is a high reactance transformer of the type used for oil-burner ignition, of approximately 10,000 V with one end grounded. One of the high tension output leads is connected to an electrode inside the vacuum chamber by means of an insulated lead-in electrode through the top of the bell jar; the other one is grounded by connecting it to the metal frame of the evaporator. The insulated upper electrode should be made of aluminium or chromium to minimize sputtering and placed some distance away from the baseplate. We used a 24-inch high bell jar with a 9-inch electrode connected through a hole at the top (Fig. 1). A glow discharge can be obtained at pressures between 500 and 10/z of Hg within the bell jar. The intensity and color of the discharge vary according to the gas pressure and the chemical composition of the residual gas. While the color is more intense at about 50-100/~, it has been shown that the efficiency of the glow discharge for making surfaces more wettable is higher at lower pressures (7). In our work the grids have been subjected to the glow discharge at a vacuum of 15 # for periods ranging from 2 to 5 minutes. To evacuate the bell jar we have used a Welch two-stage mechanical pump Model No. 1402 with the exhaust open and without the aid of a diffusion pump. With this system, the pressure falls quickly to about 15 # and it can be maintained at about that level for the 2-5 minutes necessary for the action of the glow discharge. Insufficient treatment results in incomplete spread of the droplets in a too thick layer, and excessive treatment results in spread of the stain over the entire surface of the supporting film in a layer too thin to provide enough contrast to visualize the particles, with loss of the outline of the individual droplets. At the end of the exposure time of the films to the glow discharge, the high tension supply is turned off, the valve connecting the bell jar to the pump is closed and air is admitted to the system immediately. It should be emphasized that while successful results can be obtained with carbon films evaporated even weeks before ion bombardment, spraying of the negatively stained preparation on them should be made within less than half an hour after bombardment to prevent atmospheric contamination of the surfaces. Negative staining solutions. Potassium phosphotungstate (KPT) was obtained by neutralization of a solution of phosphotungstic acid with normal K O H drop by drop until the desired pH was obtained. For routine work KPT solutions having a pH of 6.5-6.8 were used, but several experiments were also performed with solutions having a pH ranging from 5.8 to 7.2. A concentration of KPT of 2 % was found adequate when one volume of particle suspension was mixed with four volumes of KPT. Although most of the work was done with KPT, a few experiments using sodium phosphotungstate (NaPT) prepared in the same manner were also performed. Other substances were also used as negative stains: uranyl acetate was employed at a concentration of 2 %, both unbuffered with a pH of 4.2 and also neutralized with NaOH to pH 7.0 in the presence of ethylenediamine tetraacetate (EDTA) to prevent precipitate. Lead acetate was used at a concentration of 9 % with a pH of 6.1 and also alkalinized to a pH of 11.5 as described by Watson (20). Lead citrate having a pH of 12.1 was prepared as described by Reynolds (19).
PROTEIN-FREE SUSPENSIONS BY NEGATIVESTAINING
109
FIo. 1. Vacuum evaporator modified for ion bombardment. T, high reactance transformer; L, lead-in electrode. Arrow points to the baseplate where the grids are placed for ion bombardment. The suspension to be observed should be deposited on the supporting film with a low velocity spraying system (Fig. 2), after mixing with the negative staining solution at an appropriate concentration. The treated films are too fragile to withstand the use of macrodrops deposited with a pipette or a loop. A glass Vaponefrin nebulizer (Model No. 4669, obtained from the Vaponefrin Co., Edison, New Jersey) is used; it is fitted with an L-shaped tube 20 mm in diameter and ending in a wide funnel 90 mm high and 60 m m wide at the bottom (Fig. 2). The lower edge of the funnel has on occasions been scalloped to accelerate drying of the sprayed droplets, but similar results have been obtained with a funnel having a smooth, circular edge. The standard rubber bulb of the nebulizer is fitted at its free end with a plastic drying tube filled with magnesium perchlorate. In our apparatus, passage of the air through this tube has given more regular and better spread droplets, possibly by controlling the air flow. The glass slides with the grids taped on them are placed one at a time on a warming plate with a surface temperature of 26°-28°C. The spraying funnel with the nebulizer attached to it is then placed on the slide, the bulb is squirted two or three times, the funnel is rapidly
110
REISSIG AND ORRELL
f
FIG. 2. Diagram of the spraying apparatus described in the text, positioned over the grids to be sprayed.
lifted to allow drying of the droplets, and the procedure is repeated about three more times. If the rubber bulb is squirted too many times without lifting the funnel, the drops dry too slowly and artifacts appear. These consist in irregular retraction of the edges of the droplets and crystallization of the KPT. The best spraying conditions will have to be determined empirically for each apparatus. Not all the trials are successful, but by making small changes in the time of ionic bombardment or in the spraying technique, good results can be achieved. Since rapid drying is necessary for the obtention of good drops, the technique will not be successful if the humidity of the room is above 60 %. The preparations were examined with a Siemens Elmiskop I electron microscope at 80 kV with double condenser illumination. A 200-# condenser aperture and a 50-~ objective aperture were used. RESULTS This method has been applied to the study of glycogen molecules in protein-free suspensions. The preparations obtained are of high quality and the thinness and uniformity of the phosphotungstate layer prevents that the smallest molecules escape detection buried in a thick drop. Polydisperse glycogens prepared by the cold water method of Orrell and Bueding (16) have been studied (17). Fig. 4 shows three proteinfree drops of a glycogen-latex-KPT mixture prepared by conventional spraying techniques, and Fig. 3 shows a similar droplet sprayed on a film treated by our method. The magnification is the same in both pictures. The molecules appear invisible
PROTEIN-FREE SUSPENSIONS BY NEGATIVE STAINING
111
in the thick drop of the conventional preparation; they are clearly outlined, however, in the thinner droplet sprayed on the ion-bombarded film. Fig. 5 illustrates the crystallization artifacts that may be produced when the drops dry slowly in a high humidity atmosphere. This is an extreme case of total crystallization of the KPT; the most common artifact is the presence of isolated crystals within the homogeneous KPT matrix. Similar crystals may appear in droplets that have dried homogeneously, if they are stored in a damp environment. When stored in a desiccator, the preparations have been preserved in perfect condition for over a year. If a homogeneous background is desired when KPT or NaPT are the negative stains, the suspending medium for the particles should not be phosphate buffer. Phosphotungstate-phosphate mixtures appear lumpy by electron microscopy. The appearance of molecules of polydisperse glycogen from rat liver, Ascaris lumbricoides, and Schistosoma mansoni after negative staining with KPT can be compared in Figs. 6, 7, and 8. A similar picture is obtained with NaPT and with uranyl acetate. An advantage of this method is that unconventional substances, which could either react with the added protein or be very sensitive to the heat effects of the electron beam in the large, thick drops conventionally used, can be employed as negative stains. Such is the case of the lead salts at alkaline pH that are generally used for the staining of glycogen in sections (18). Fig. 9 shows pure glycogen molecules from Ascaris Iumbricoides prepared by the method of Orrell and Bueding (16) which have been mixed with Reynold's lead citrate (19) 10 minutes before spraying. A carbon dioxide absorbent (Ascarite, A. H. Thomas) was used to fill the drying tube connected to the rubber bulb used for spraying. Under the conditions of rapid drying described under Materials and Methods, this has been sufficient to prevent the appearance of lead precipitates. The lead salts granulate slightly under the electron beam, but by rapid photography of unexamined fields under double condenser illumination, heat damage can be reduced to a minimum. Similar results have been obtained with Watson's lead acetate (20) and with nonalkalinized lead acetate. DISCUSSION The preservation of structural detail obtained by this technique is equal, if not superior, to that obtained by conventional methods. The uniformity of distribution of the particles and the possibility of evaluating whole drop patterns make this method particularly suitable for the evaluation of particle size distribution, particle counts and molecular weight determinations by the method of Williams and Backus (21), as shown in Fig. 3. The main drawback of these ion-bombarded films is their fragility, even greater
112
REISSIG AND ORRELL
than that of untreated carbon films, which makes them suitable only for use with low velocity spraying techniques; they break when immersed in liquids or when large drops are deposited on them with a pipette. It is possible to treat by this method Formvar-carbon films, which are suitable for supporting large drops deposited on them with a pipette, but even these have to be handled with extreme care due to their tendency to break. . : Attempts to use this principle for the preparation of wettable supporting films had previously been made by Huxley and Zubay without success (11): O u r data show that good results can be obtained by the procedure just described. However, some of the critical variables, such a s the dosage of ionic b o m b a r d m e n t needed for correct spreading and the rate of drying of the drops, are impossible to measure and control adequately. Although a good proportion of the trials are successful, not every batch of sprayed grids is equally good. When uniformly spread drops with perfect edges are needed, as is the case for particle counting procedures, it may be .necessary to make several trials with minor adjustments of the ionic b o m b a r d m e n t time or the spraying technique. It should be noted that the p H of the negative staining solution apparently did not have any effect on the spreading of the droplets; while the K P T solution used for routine work had a p H of 6.5-6.8, similar results have been obtained with K P T at a p H ranging from 5.8 to 7.2, with NaPT, with uranyl acetate unbuffered or buffered in the presence of EDTA, and with various lead salts ranging in p H from 6.1 to 12.1. Although well spread, the mixtures containing uranyl acetate and E D T A were consistently very poor in contrast when observed with the electron microscope. On the other hand, small changes in the temperature of the surface on which the films were placed for spraying or in the rate of drying of the droplets, which could be empirically controlled by the number of times the rubber bulb was squeezed before lifting the funnel or by the promptness in lifting the spraying apparatus away from the films, appeared to have a great influence on droplet shape and edge retraction. This last statement is intended only as a practical working guide, since a quantitative evaluation of the factors involved in drop spreading is extremely difficult and has not been performed.
FIG. 3. Electron micrograph of a microdrop Containing glycogen molecules, polystyrene latex spheres 88 nm in diameter, and KPT, sprayed on an ion-bombarded film. Such drops can be used for particle counting techniques, x 11,000. FIG. 4. Three drops of the preparation shown in Fig. 3, but sprayed onto a conventional, untreated carbon film. The magnification on both figures is identical. The glycogen and the latex spheres cannot be seen because of the thickness of the drop. FIG. 5. Drop from the preparation shown in Figs. 3 and 4, but sprayed on an ion-bombarded film in a room where the humidity was over 90 %. All the KPT has dried in the form of crystals, making recognition of the molecules impossible, x 3800.
ir~
i
i
I I
I
' ! (iiiii~((i; :i i¸¸!i]!i; !ii!iiiilj!!~i~
114
REISSIG AND ORRELL
PROTEIN-FREE SUSPENSIONS BY NEGATIVE STAINING
115
FIG. 9. Field from a preparation of glycogen from Ascaris lumbricoides mixed with Reynold's lead citrate (pH 12.1) and sprayed 10 minutes after mixing. Note that the glycogen does not appear stained by the lead salts. × 140,000.
The exact m e c h a n i s m b y which the glow discharge acts in m a k i n g films m o r e wettable has n o t been d e t e r m i n e d . A s stated b y H o l l a n d (5): " T h e b o m b a r d m e n t process is very c o m p l e x a n d it is n o t surprising t h a t a satisfactory single t h e o r y of the cleaning m e c h a n i s m has n o t been a d v a n c e d . " H i g h voltage glow discharges with characteristics similar to those e m p l o y e d b y us (where the gas pressure was a b o u t 10-20 # for a n a p p l i e d voltage of the o r d e r of 5-10 kV) have been used b y m a n y w o r k e r s to clean surfaces p r i o r to film d e p o s i t i o n (5). R e m o v a l of surface a d s o r b e d layers is p r o b a b l y of m a j o r i m p o r t a n c e in this process. H o w e v e r , an oxide film c o u l d be s p u t t e r e d at the s a m e time f r o m the discharge electrodes (6) or f r o m silicone residual v a p o r s , a n d if FIG. 6. Field from a preparation of rat liver glycogen mixed with KPT. x 130,000. Fio. 7. Field from a preparation of glycogen from Ascaris lurnbricoides, x 200,000. FIo. 8. Field from a preparation of glycogen from Schistosoma mansoni mixed with KPT. Very few of the large rosettes seen in Figs. 6 and 7 are found; most of the molecules are of the small, single subunit type. x 80,000.
116
REISSIG AND ORRELL
silica films were formed, they would be very difficult to detect since in all probability they would be wettable (13). Two lines of evidence support the view that removal of layers may be the main phenomenon occurring in our system: first, if a shiny aluminium plate is lightly coated with an evaporated carbon film and then exposed for a period of several hours to the glow discharge under the conditions described in the preceeding section for the bombardment of supporting films, the brownish carbon coating is removed, exposing the silvery metal surface. Also carbon supporting films exposed to the glow discharge for prolonged periods are extremely fragile and exhibit an irregular, wrinkled surface when viewed in the electron microscope after shadowcasting. It is well established, however, that thin films can be deposited in a glow discharge (8, 12). Konig and Helwig (14) have used it for the deposition of polymer films from hydrocarbon vapors admitted to the ionization chamber. While it is impossible to rule out that a deposition process occurs to any extent in our system, the evidence given above plus the fact that the conditions under which a discharge was produced in the system described by Konig and Helwig concerning applied voltage, composition, and pressure of the residual gas and electrode geometry were very different from ours, give support to the view that in our system removal of adsorbed layers or etching of the film surface are probably the main factors involved in the cleaning process. The results obtained when glycogen was mixed with the alkaline lead salts currently used for the staining of glycogen in sections have been unexpected. They are quite reproducible; they have been consistently obtained with a variety of glycogens, extracted and purified from different sources. However, when fixed and embedded sections of the same tissues from which the glycogens were extracted were stained with the same batch of alkaline lead salts used for negative staining, the glycogen areas in the sections appeared positively stained, while the purified glycogens in the droplets did not react with the lead. This lack of stainability of the isolated glycogen molecules cannot be considered a contrast phenomenon of the kind described by Bradley (1). If a glycogen preparation a hundred times more concentrated than the usual ones is mixed with the lead stain, allowed to interact with it for a period ranging from 1 to 30 minutes, rapidly diluted in distilled water 100-fold to reduce the intensity of the staining solution maintaining the particle concentration at the usual level and if this dilute suspension is immediately sprayed, no positively stained molecules can be seen. The molecules are still present, as can be demonstrated by shadowcasting techniques. Thus, purified isolated glycogen does not appear to react with alkaline lead salts. The significance of these results is not known and is currently under study. This work was supported by grants AI-07806, HD-02367, and AI-08022 from the U.S. Public Health Service and by a grant from the John A. Hartford Foundation. The authors wish to thank Dr. Michael Beer for his valuable suggestions.
PROTEIN-FREE SUSPENSIONSBY NEGATIVESTAINING
117
REFERENCES
1. BRADLEY,D. E., J. Gen. Microbiol. 29, 508 (1962). 2. BRENNER, S. and HORNE, R. W., Biophys. Biochim. Acta 34, 103 (1959). 3. HALL, C. E., J. Biophys. Biochem. Cytol. 2, 625 (1956). 4. HOLLAND,L., Vacuum Deposition of Thin Films, p. 74. Wiley, New York, 1961. 5. - ibid, p. 75. 6. - ibid. p. 76. 7. - - - - ibid, pp. 85 and 87. 8. - ibid, p. 88. 9. HORNZ, R. W., in KAY, D. (Ed.), Techniques for Electron Microscopy, 2nd Ed., p. 330. Davis, Philadelphia, 1965. 10. - - - - ibid, p. 331. 11. HUXLEY, C. E. and ZUBAY, G., J. Mol. Biol. 2, 10 (1960). 12. KAY, E., Advan. Electron. Electron Phys. 17, 297 (1962). 13. - ibid, p. 301. 14. KONm, H. and HELWm, G., Z. Phys. 129, 491 (1951). 15. MORDOn, J., LELOm, L. F. and KRISMAN, C. R., Proe. Natl. Acad. Sci.. U.S. 53, 86 (1965). 16. ORRELL, S. A. and BUEDING, E., 3". Biol. Chem. 239, 4021 (1964). 17. REISSIG,M. and ORRELL, S. A., J. Appl. Phys. 34, 2510 (1963). 18. REVEL, J. P., J. Histoehem. Cytochem. 12, 104 (1964). 19. REYNOLDS,E. S., J. Cell Biol. 17, 208 (1963). 20. WATSON, M. L., J. Biophys. Bioehem. Cytol. 4, 727 (1958). 21. WILLIAMS,R. C. and BACKUS, R. C., J. Am. Chem. Soe. 71, 4052 (1949).
J. ULTRASTRUCXURERESEARCH32, 107-117 (1970)
107
A Technique for the Electron Microscopy of Protein-Free Particle Suspensions by the Negative Staining Method I MAGDALENA REISS1G AND NTANLEYA. ORRELL
Department of Pathobiology, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, Maryland 21205 Received August 21, 1969, and in revised form January 5, 1970 A method has been developed for the electron microscopic study of proteinfree particle suspensions by the negative staining method without the addition of substances that would lower the surface tension of the preparation to facilitate spreading. It involves treating the carbon supporting films with a high voltage glow discharge under carefully controlled conditions, before spraying on them the suspensions to be observed. Its applications include particle counts in microdrops with the addition of latex standards and evaluation of particle size distributions. It is particularly useful in those cases when one wishes to avoid the use of additives, such as detergents or soluble proteins, which may react with the particles under study or with the negative staining solution. Use of the negative staining method (2) for the study of particulate materials by electron microscopy, requires that the preparation be spread on the supporting film in a thin layer (9). Since the aqueous solutions of phosphotungstate and other negative stains generally used do not wet adequately the supporting films, the method has been successful only when the surface tension of the particle suspension was low enough to permit adequate spreading. Most biological preparations contain some soluble protein, either as the substance under study or as a contaminating impurity, which act as a surfactant. If such an agent is not present, either albumin (10) or detergents (15) can be added for this purpose. Another approach to this problem would be to devise a method to modify the surface of the supporting film in such a way that it would be readily wettable by the preparation, even if its surface tension was high. The present report is concerned with the description of such a technique, which was devised for the study of protein-free suspensions of glycogen. It is based on the principle, known for some time, that glass and metal surfaces can be cleaned and made surface active by exposing them to a high voltage glow discharge passed between metal electrodes in a continuously exhausted vessel (4). i This investigation was supported in part by the United States-Japan Cooperative Medical Science Program administered by the National Institute of Allergy and Infectious Diseases of The National Institutes of Health, Department of Health, Education, and Welfare.
108
REISSIG AND ORRELL
MATERIALS A N D METHODS The following procedure was developed for preparing wettable carbon films: pure carbon films evaporated on freshly cleaved mica by the method of Hall (3) are mounted on copper grids and fastened with tape to glass slides. They are then exposed to a high voltage glow discharge in a standard vacuum evaporator with a modified bell jar and a power supply for the glow discharge. The power supply for the discharge is a high reactance transformer of the type used for oil-burner ignition, of approximately 10,000 V with one end grounded. One of the high tension output leads is connected to an electrode inside the vacuum chamber by means of an insulated lead-in electrode through the top of the bell jar; the other one is grounded by connecting it to the metal frame of the evaporator. The insulated upper electrode should be made of aluminium or chromium to minimize sputtering and placed some distance away from the baseplate. We used a 24-inch high bell jar with a 9-inch electrode connected through a hole at the top (Fig. 1). A glow discharge can be obtained at pressures between 500 and 10/z of Hg within the bell jar. The intensity and color of the discharge vary according to the gas pressure and the chemical composition of the residual gas. While the color is more intense at about 50-100/~, it has been shown that the efficiency of the glow discharge for making surfaces more wettable is higher at lower pressures (7). In our work the grids have been subjected to the glow discharge at a vacuum of 15 # for periods ranging from 2 to 5 minutes. To evacuate the bell jar we have used a Welch two-stage mechanical pump Model No. 1402 with the exhaust open and without the aid of a diffusion pump. With this system, the pressure falls quickly to about 15 # and it can be maintained at about that level for the 2-5 minutes necessary for the action of the glow discharge. Insufficient treatment results in incomplete spread of the droplets in a too thick layer, and excessive treatment results in spread of the stain over the entire surface of the supporting film in a layer too thin to provide enough contrast to visualize the particles, with loss of the outline of the individual droplets. At the end of the exposure time of the films to the glow discharge, the high tension supply is turned off, the valve connecting the bell jar to the pump is closed and air is admitted to the system immediately. It should be emphasized that while successful results can be obtained with carbon films evaporated even weeks before ion bombardment, spraying of the negatively stained preparation on them should be made within less than half an hour after bombardment to prevent atmospheric contamination of the surfaces. Negative staining solutions. Potassium phosphotungstate (KPT) was obtained by neutralization of a solution of phosphotungstic acid with normal K O H drop by drop until the desired pH was obtained. For routine work KPT solutions having a pH of 6.5-6.8 were used, but several experiments were also performed with solutions having a pH ranging from 5.8 to 7.2. A concentration of KPT of 2 % was found adequate when one volume of particle suspension was mixed with four volumes of KPT. Although most of the work was done with KPT, a few experiments using sodium phosphotungstate (NaPT) prepared in the same manner were also performed. Other substances were also used as negative stains: uranyl acetate was employed at a concentration of 2 %, both unbuffered with a pH of 4.2 and also neutralized with NaOH to pH 7.0 in the presence of ethylenediamine tetraacetate (EDTA) to prevent precipitate. Lead acetate was used at a concentration of 9 % with a pH of 6.1 and also alkalinized to a pH of 11.5 as described by Watson (20). Lead citrate having a pH of 12.1 was prepared as described by Reynolds (19).
PROTEIN-FREE SUSPENSIONS BY NEGATIVESTAINING
109
FIo. 1. Vacuum evaporator modified for ion bombardment. T, high reactance transformer; L, lead-in electrode. Arrow points to the baseplate where the grids are placed for ion bombardment. The suspension to be observed should be deposited on the supporting film with a low velocity spraying system (Fig. 2), after mixing with the negative staining solution at an appropriate concentration. The treated films are too fragile to withstand the use of macrodrops deposited with a pipette or a loop. A glass Vaponefrin nebulizer (Model No. 4669, obtained from the Vaponefrin Co., Edison, New Jersey) is used; it is fitted with an L-shaped tube 20 mm in diameter and ending in a wide funnel 90 mm high and 60 m m wide at the bottom (Fig. 2). The lower edge of the funnel has on occasions been scalloped to accelerate drying of the sprayed droplets, but similar results have been obtained with a funnel having a smooth, circular edge. The standard rubber bulb of the nebulizer is fitted at its free end with a plastic drying tube filled with magnesium perchlorate. In our apparatus, passage of the air through this tube has given more regular and better spread droplets, possibly by controlling the air flow. The glass slides with the grids taped on them are placed one at a time on a warming plate with a surface temperature of 26°-28°C. The spraying funnel with the nebulizer attached to it is then placed on the slide, the bulb is squirted two or three times, the funnel is rapidly
110
REISSIG AND ORRELL
f
FIG. 2. Diagram of the spraying apparatus described in the text, positioned over the grids to be sprayed.
lifted to allow drying of the droplets, and the procedure is repeated about three more times. If the rubber bulb is squirted too many times without lifting the funnel, the drops dry too slowly and artifacts appear. These consist in irregular retraction of the edges of the droplets and crystallization of the KPT. The best spraying conditions will have to be determined empirically for each apparatus. Not all the trials are successful, but by making small changes in the time of ionic bombardment or in the spraying technique, good results can be achieved. Since rapid drying is necessary for the obtention of good drops, the technique will not be successful if the humidity of the room is above 60 %. The preparations were examined with a Siemens Elmiskop I electron microscope at 80 kV with double condenser illumination. A 200-# condenser aperture and a 50-~ objective aperture were used. RESULTS This method has been applied to the study of glycogen molecules in protein-free suspensions. The preparations obtained are of high quality and the thinness and uniformity of the phosphotungstate layer prevents that the smallest molecules escape detection buried in a thick drop. Polydisperse glycogens prepared by the cold water method of Orrell and Bueding (16) have been studied (17). Fig. 4 shows three proteinfree drops of a glycogen-latex-KPT mixture prepared by conventional spraying techniques, and Fig. 3 shows a similar droplet sprayed on a film treated by our method. The magnification is the same in both pictures. The molecules appear invisible
PROTEIN-FREE SUSPENSIONS BY NEGATIVE STAINING
111
in the thick drop of the conventional preparation; they are clearly outlined, however, in the thinner droplet sprayed on the ion-bombarded film. Fig. 5 illustrates the crystallization artifacts that may be produced when the drops dry slowly in a high humidity atmosphere. This is an extreme case of total crystallization of the KPT; the most common artifact is the presence of isolated crystals within the homogeneous KPT matrix. Similar crystals may appear in droplets that have dried homogeneously, if they are stored in a damp environment. When stored in a desiccator, the preparations have been preserved in perfect condition for over a year. If a homogeneous background is desired when KPT or NaPT are the negative stains, the suspending medium for the particles should not be phosphate buffer. Phosphotungstate-phosphate mixtures appear lumpy by electron microscopy. The appearance of molecules of polydisperse glycogen from rat liver, Ascaris lumbricoides, and Schistosoma mansoni after negative staining with KPT can be compared in Figs. 6, 7, and 8. A similar picture is obtained with NaPT and with uranyl acetate. An advantage of this method is that unconventional substances, which could either react with the added protein or be very sensitive to the heat effects of the electron beam in the large, thick drops conventionally used, can be employed as negative stains. Such is the case of the lead salts at alkaline pH that are generally used for the staining of glycogen in sections (18). Fig. 9 shows pure glycogen molecules from Ascaris Iumbricoides prepared by the method of Orrell and Bueding (16) which have been mixed with Reynold's lead citrate (19) 10 minutes before spraying. A carbon dioxide absorbent (Ascarite, A. H. Thomas) was used to fill the drying tube connected to the rubber bulb used for spraying. Under the conditions of rapid drying described under Materials and Methods, this has been sufficient to prevent the appearance of lead precipitates. The lead salts granulate slightly under the electron beam, but by rapid photography of unexamined fields under double condenser illumination, heat damage can be reduced to a minimum. Similar results have been obtained with Watson's lead acetate (20) and with nonalkalinized lead acetate. DISCUSSION The preservation of structural detail obtained by this technique is equal, if not superior, to that obtained by conventional methods. The uniformity of distribution of the particles and the possibility of evaluating whole drop patterns make this method particularly suitable for the evaluation of particle size distribution, particle counts and molecular weight determinations by the method of Williams and Backus (21), as shown in Fig. 3. The main drawback of these ion-bombarded films is their fragility, even greater
112
REISSIG AND ORRELL
than that of untreated carbon films, which makes them suitable only for use with low velocity spraying techniques; they break when immersed in liquids or when large drops are deposited on them with a pipette. It is possible to treat by this method Formvar-carbon films, which are suitable for supporting large drops deposited on them with a pipette, but even these have to be handled with extreme care due to their tendency to break. . : Attempts to use this principle for the preparation of wettable supporting films had previously been made by Huxley and Zubay without success (11): O u r data show that good results can be obtained by the procedure just described. However, some of the critical variables, such a s the dosage of ionic b o m b a r d m e n t needed for correct spreading and the rate of drying of the drops, are impossible to measure and control adequately. Although a good proportion of the trials are successful, not every batch of sprayed grids is equally good. When uniformly spread drops with perfect edges are needed, as is the case for particle counting procedures, it may be .necessary to make several trials with minor adjustments of the ionic b o m b a r d m e n t time or the spraying technique. It should be noted that the p H of the negative staining solution apparently did not have any effect on the spreading of the droplets; while the K P T solution used for routine work had a p H of 6.5-6.8, similar results have been obtained with K P T at a p H ranging from 5.8 to 7.2, with NaPT, with uranyl acetate unbuffered or buffered in the presence of EDTA, and with various lead salts ranging in p H from 6.1 to 12.1. Although well spread, the mixtures containing uranyl acetate and E D T A were consistently very poor in contrast when observed with the electron microscope. On the other hand, small changes in the temperature of the surface on which the films were placed for spraying or in the rate of drying of the droplets, which could be empirically controlled by the number of times the rubber bulb was squeezed before lifting the funnel or by the promptness in lifting the spraying apparatus away from the films, appeared to have a great influence on droplet shape and edge retraction. This last statement is intended only as a practical working guide, since a quantitative evaluation of the factors involved in drop spreading is extremely difficult and has not been performed.
FIG. 3. Electron micrograph of a microdrop Containing glycogen molecules, polystyrene latex spheres 88 nm in diameter, and KPT, sprayed on an ion-bombarded film. Such drops can be used for particle counting techniques, x 11,000. FIG. 4. Three drops of the preparation shown in Fig. 3, but sprayed onto a conventional, untreated carbon film. The magnification on both figures is identical. The glycogen and the latex spheres cannot be seen because of the thickness of the drop. FIG. 5. Drop from the preparation shown in Figs. 3 and 4, but sprayed on an ion-bombarded film in a room where the humidity was over 90 %. All the KPT has dried in the form of crystals, making recognition of the molecules impossible, x 3800.
ir~
i
i
I I
I
' ! (iiiii~((i; :i i¸¸!i]!i; !ii!iiiilj!!~i~
114
REISSIG AND ORRELL
PROTEIN-FREE SUSPENSIONS BY NEGATIVE STAINING
115
FIG. 9. Field from a preparation of glycogen from Ascaris lumbricoides mixed with Reynold's lead citrate (pH 12.1) and sprayed 10 minutes after mixing. Note that the glycogen does not appear stained by the lead salts. × 140,000.
The exact m e c h a n i s m b y which the glow discharge acts in m a k i n g films m o r e wettable has n o t been d e t e r m i n e d . A s stated b y H o l l a n d (5): " T h e b o m b a r d m e n t process is very c o m p l e x a n d it is n o t surprising t h a t a satisfactory single t h e o r y of the cleaning m e c h a n i s m has n o t been a d v a n c e d . " H i g h voltage glow discharges with characteristics similar to those e m p l o y e d b y us (where the gas pressure was a b o u t 10-20 # for a n a p p l i e d voltage of the o r d e r of 5-10 kV) have been used b y m a n y w o r k e r s to clean surfaces p r i o r to film d e p o s i t i o n (5). R e m o v a l of surface a d s o r b e d layers is p r o b a b l y of m a j o r i m p o r t a n c e in this process. H o w e v e r , an oxide film c o u l d be s p u t t e r e d at the s a m e time f r o m the discharge electrodes (6) or f r o m silicone residual v a p o r s , a n d if FIG. 6. Field from a preparation of rat liver glycogen mixed with KPT. x 130,000. Fio. 7. Field from a preparation of glycogen from Ascaris lurnbricoides, x 200,000. FIo. 8. Field from a preparation of glycogen from Schistosoma mansoni mixed with KPT. Very few of the large rosettes seen in Figs. 6 and 7 are found; most of the molecules are of the small, single subunit type. x 80,000.
116
REISSIG AND ORRELL
silica films were formed, they would be very difficult to detect since in all probability they would be wettable (13). Two lines of evidence support the view that removal of layers may be the main phenomenon occurring in our system: first, if a shiny aluminium plate is lightly coated with an evaporated carbon film and then exposed for a period of several hours to the glow discharge under the conditions described in the preceeding section for the bombardment of supporting films, the brownish carbon coating is removed, exposing the silvery metal surface. Also carbon supporting films exposed to the glow discharge for prolonged periods are extremely fragile and exhibit an irregular, wrinkled surface when viewed in the electron microscope after shadowcasting. It is well established, however, that thin films can be deposited in a glow discharge (8, 12). Konig and Helwig (14) have used it for the deposition of polymer films from hydrocarbon vapors admitted to the ionization chamber. While it is impossible to rule out that a deposition process occurs to any extent in our system, the evidence given above plus the fact that the conditions under which a discharge was produced in the system described by Konig and Helwig concerning applied voltage, composition, and pressure of the residual gas and electrode geometry were very different from ours, give support to the view that in our system removal of adsorbed layers or etching of the film surface are probably the main factors involved in the cleaning process. The results obtained when glycogen was mixed with the alkaline lead salts currently used for the staining of glycogen in sections have been unexpected. They are quite reproducible; they have been consistently obtained with a variety of glycogens, extracted and purified from different sources. However, when fixed and embedded sections of the same tissues from which the glycogens were extracted were stained with the same batch of alkaline lead salts used for negative staining, the glycogen areas in the sections appeared positively stained, while the purified glycogens in the droplets did not react with the lead. This lack of stainability of the isolated glycogen molecules cannot be considered a contrast phenomenon of the kind described by Bradley (1). If a glycogen preparation a hundred times more concentrated than the usual ones is mixed with the lead stain, allowed to interact with it for a period ranging from 1 to 30 minutes, rapidly diluted in distilled water 100-fold to reduce the intensity of the staining solution maintaining the particle concentration at the usual level and if this dilute suspension is immediately sprayed, no positively stained molecules can be seen. The molecules are still present, as can be demonstrated by shadowcasting techniques. Thus, purified isolated glycogen does not appear to react with alkaline lead salts. The significance of these results is not known and is currently under study. This work was supported by grants AI-07806, HD-02367, and AI-08022 from the U.S. Public Health Service and by a grant from the John A. Hartford Foundation. The authors wish to thank Dr. Michael Beer for his valuable suggestions.
PROTEIN-FREE SUSPENSIONSBY NEGATIVESTAINING
117
REFERENCES
1. BRADLEY,D. E., J. Gen. Microbiol. 29, 508 (1962). 2. BRENNER, S. and HORNE, R. W., Biophys. Biochim. Acta 34, 103 (1959). 3. HALL, C. E., J. Biophys. Biochem. Cytol. 2, 625 (1956). 4. HOLLAND,L., Vacuum Deposition of Thin Films, p. 74. Wiley, New York, 1961. 5. - ibid, p. 75. 6. - ibid. p. 76. 7. - - - - ibid, pp. 85 and 87. 8. - ibid, p. 88. 9. HORNZ, R. W., in KAY, D. (Ed.), Techniques for Electron Microscopy, 2nd Ed., p. 330. Davis, Philadelphia, 1965. 10. - - - - ibid, p. 331. 11. HUXLEY, C. E. and ZUBAY, G., J. Mol. Biol. 2, 10 (1960). 12. KAY, E., Advan. Electron. Electron Phys. 17, 297 (1962). 13. - ibid, p. 301. 14. KONm, H. and HELWm, G., Z. Phys. 129, 491 (1951). 15. MORDOn, J., LELOm, L. F. and KRISMAN, C. R., Proe. Natl. Acad. Sci.. U.S. 53, 86 (1965). 16. ORRELL, S. A. and BUEDING, E., 3". Biol. Chem. 239, 4021 (1964). 17. REISSIG,M. and ORRELL, S. A., J. Appl. Phys. 34, 2510 (1963). 18. REVEL, J. P., J. Histoehem. Cytochem. 12, 104 (1964). 19. REYNOLDS,E. S., J. Cell Biol. 17, 208 (1963). 20. WATSON, M. L., J. Biophys. Bioehem. Cytol. 4, 727 (1958). 21. WILLIAMS,R. C. and BACKUS, R. C., J. Am. Chem. Soe. 71, 4052 (1949).