JOURNAL OF ELECTRON MICROSCOPY TECHNIQUE 18:183-191 (1991)
Preservation of EDTA-Expanded Grid-Mounted Chromosomes and Nuclei for Electron Microscopy Using a Specially Designed Freeze-Dryer PHILIP S. WOODS, MYRON C. LEDBETTER, AND NEAL TEMPEL Biology Department, Brookhaven National Laboratory, Upton, New York 11973
KEY WORDS
Whole-mounts, Freeze-dry, Freeze-substitution, TEM
ABSTRACT We describe methods for freezing and drying EDTA-expanded, fixed metaphase chromosomes and nuclei, attached to grids as whole-mounts, for transmission electron microscopy. These methods use a special apparatus that is simple to construct. While separate freezers and dryers are commercially available, one for freezing blocks of tissue by slamming them against a cold metal surface, and the other for vacuum drying the frozen tissue, our apparatus is designed for gentler, cryogenic liquid plunge freezing and drying, sequentially, in the same apparatus, thus avoiding any compression or damage to the sepcimen. Use of a cryoprotectant is not essential; however, good results are obtained more often when 20% ethanol is used. Freezing is accomplished by rapid propulsion of the grid, with specimens attached, into slushy N, (-210°C) within the drying chamber; drying is automatic, by either sublimation under vacuum or by solvent substitution using absolute ethanol followed by acetone, which, in turn, is removed with a critical-point dryer. The apparatus offers a means of drying chromosomes and nuclei in a n expanded state, and avoids the shrinkage of these structures that occurs during stepwise passage through increasing concentrations of ethanol or acetone. INTRODUCTION
MATERIALS AND METHODS The apparatus When isolated formalin-fixed metaphase chromoFigure 1 shows our apparatus and the inset shows a somes are examined as whole-mounts by the 100 kv disassembled view of the electric heater, and the transmission electron microscope, they are nearly sleeve/cover assembly, the latter serving as a wateropaque to the electron beam (Woods et al., 1977) and vapor condensing trap located in the block. Figure 2 is reveal little of their internal structure. We have at- a sectional in-scale diagram outlining the major comtempted to render these chromosomes more accessible ponents of the apparatus within the vacuum chamber. to electron microscopy by expanding them with the This design uses a modified 9-port freeze-etch collar chelating agent Na,EDTA.2H20 (Woods et al., 1979). from the D F E S apparatus (Denton Vacuum, Inc., This treatment reduced their density; however, subse- Cherry Hill, NJ). We mount the collar separately on a n quent dehydration with acetone or ethanol always aluminum base plate. The collar is evacuated with a caused the chromosomes to recondense again. Clearly, rotary vacuum pump using a liquid Nz trap to protect another method of drying was needed to preserve the the specimen from backstreaming of oil. With this modchromosomes in a n open state. Freeze-drying was con- est apparatus we have obtained usable specimens a templated. Several dryers had been either described or majority of the time. were available commercially, e.g., ones by Chiovetti et al. (1987)) Coulter and Terracio (19781, Edelmann Freezing the specimen (1978)) Elder et al. (1986), and Linner e t al. (1986); however, these seemed inappropriate for grid-mounted 1. The liquid cryogen used is slushy N2 (-210°C) specimens, since they usually were used with “slammade by first filling with liquid N, the Dewar flask (I), frozen” blocks of tissue for eventual embedding and sectioning (Phillips and Boyne, 1984). Pawley and Ris (1987) used a commercial slamming device which they modified with a cushion for freezing cells growing on grids; the grids then were transferred to a dryer of their Received October 4, 1 9 8 9 accepted in revised form August 25, 1990. design. Our apparatus is simple to construct, and proAddress reprint requests to Dr. Philip S. Woods, Biology Department, vides for gentle “plunge-freezing” and subsequent dry- Brookhaven National Laboratory, Upton, NY 11973. ing in the same apparatus, avoiding compression of the Philip S.Woods is a guest scientist of the Biology Department, Brookhaven specimen and damage to it during transfer. This paper National Laboratory, Upton, NY 11973. Abbreviations: FD = freeze-dried, meaning frozen from either water or 20% describes the apparatus and its use, especially in the ethanol and dried by vacuum sublimation; FS = freeze-substitution, meaning study of EDTA-expanded, fixed chromosomes and nu- frozen from either water or 20% ethanol and dried by ethanol substitution and use of a critical-point dryer and acetone. clei isolated from growing root tips of a higher plant.
PUBLISHED 1991 WILEY-LISS, INC.
184
P.S. WOODS ET AL.
Fig. 1. The freeze-dry apparatus consisting of: (A) Denton freezeetch 9-port collar, (BJ aluminum base plate, (C) optical meter relay, double set-point, thermocouple, temperature gauge, (D) direct current power supply, (EJ dry-ice chilled Drierite water-vapor trap, (F) inlet hose from high pressure N, tank, (GI inlet valve to chamber for dry N, gas, (H) main valve to chamber from mechanical pump, (I) Dewar flask, (J)heat absorber with copper post, (K) movable arm for lifting and lowering heat absorber and post, (L) block thermocouple wire, (M) heater thermocouple wire, (N) Copper-Constantan thermocouple
plugs, (0)plastic beaker of P,O,, (P) Lucite cover to port for electric heater and thermocouple wires. Inset: Shows the electric heatersleeveicover assembly consisting of: (A) the heater (its plastic screen fence is removed), two insulated stainless steel electrical lead wires, and the heater thermocouple wire, that pass through holes in the top of the sleeveicover (B).When the heater is enclosed within the sleeve/ cover, and the two, in turn, are fitted into the large well of the block (as shown in Fig. 21, the heater does not touch the sleevdcover nor does it touch the block except by the needle legs.
containing the 3 kg copper block and heater-sleeve/ cover assembly. After nucleate boiling subsides, the chamber is evacuated with a rotary pump (112 hp, Duo Seal, 160 liters per minute capacity) until the nitrogen is almost completely frozen. The temperature of the block should read -217"C, about 2" warmer than the lowest temperature attainable. The chamber then is brought to atmospheric pressure within 20 seconds with dry N, gas from a high pressure cylinder (not shown) equipped with a pressure reducer, using the small inlet valve (G) on the collar. This gas is cooled and dried a s it passes through the dry-ice chilled Drierite (anhydrous CaSO,) water-vapor trap (E). 2. Grids are picked up with forceps from ice-chilled 20% ethanol and placed, specimen side up, on the end of a long thin wooden stick, blotted briefly at one edge, and together turned and snapped firmly by bending and releasing the end of the stick over the hole in the glass plate. The grid is propelled a t high velocity into the slush. It is important that freezing of the grids
begins within the first few seconds after atmospheric pressure is reached before the temperature a t the surface of the slush becomes too high. A plastic screen fence (Fig. 21, wrapped around the block, prevents the grids from falling under the block. A CopperConstantan thermocouple is mounted in the block to monitor the temperature of the block with its wire passing through a port in the collar connecting by way of wire (L) and a set of Copper-Constantan thermocouple plugs (N) to a double (or single) set- point, thermocouple, temperature gauge (C) made by A.P.I. Instrument Co., Clinton, MA. 3. The remaining liquid N, is boiled away by raising the block temperature. This is done by attaching the post of the heat absorber (J)to the heat conductor on the block (Fig. 2) using the movable arm (K) and heating the absorber with a heat lamp (not shown). A piece of lead is fused to the corner (underside) of the heat absorber opposite its post (Fig. 2) to serve as a counterbalance.
FREEZE-DRYING GRID-MOUNTED SPECIMENS
185
Fig. 2. Diagrammatic view of the part of the apparatus in the vacuum chamber.
Drying the specimens-the vacuum sublimation method (FD)
spectively, the electric heater is turned on to dry the specimens. This is accomplished as follows: the switch on the direct-current power supply (2 amps, 10 volts 1. Grids are pushed through the holes of the sleeve1 capacity) is turned on. Current to the embedded resiscover (Fig. 1, inset and 2 ) where they fall directly onto tor in the heater (Fig. 1, inset) is regulated by the setthe electric heater, which consists of a 30 ohm, 5 watt point temperature gauge. For this the heater thermoresistor embedded in a small copper cylinder with sil- couple is connected to the gauge and the set-pointer on icone rubber adhesive, and thermally isolated from the the gauge is moved to a suitable drying temperature. block with three needle-sharp stainless steel legs. An- We find that -75°C yields good results. The amount of other smaller plastic screen (not shown) is wrapped heat is controlled by the voltage of the power supply, around the heater to prevent grids from falling under and the set-pointer determines the temperature a t it. There is another Copper-Constantan thermocouple which the heater will be maintained. We find that 5 mounted in the heater with its wire passing through volts raises the temperature of the heater from - 135°C another port on the collar (P) to the wire (M) that may to -75°C in about 2 hours. The temperature of the be connected to the same temperature gauge (C) by block and sleeveicover rises a t a much slower rate and another Copper-Constantan thermocouple plug (N). both are always colder than the heater. The set-pointer Exchange of these plugs between the two thermocou- gauge maintains the heater constantly a t the chosen ples allows the temperature of the block or heater to be temperature of -75°C until the block later approaches read by the same temperature gauge, provided the to- -75"C, at which time the heater automatically shuts tal lengths of each thermocouple wire are equal. After off. During drying the colder sleeveicover traps water the grids are placed on the heater, a small beaker of the molecules as they sublime from the grids and the specdesiccant P,O, (0)is passed through the hole in the imens gradually dry. The cycling time is 1 to 2 hours glass plate and placed on the block. When all the liquid when the temperature is held at -75°C. Denton VacN, is gone, the heat lamp is turned off and the heat uum, Inc. informs us (personal communication) that absorber with post is disconnected. the rate of sublimation of ice is approximately 10 pm 2 . The hole in the glass plate is covered with a plug per minute at -75"C, under optimum conditions of vacand the chamber is evacuated. The pump is left on uum and vapor pressure. Since our specimens are less during the entire drying period. In addition to the liq- than 100 pm in thickness (including the ice over the uid N2 trap on the vacuum line to the pump, there is an expanded nuclei), they are dry within the first 10 minoptic-dense foreline trap made by Ion Equipment Corp., utes if these data hold for the conditions in our appaSanta Clara, CA (both traps not shown). ratus. 3. After 5-10 minutes, when the vacuum comes 4. By the afternoon of the next day the heater and close t o its best value of 1 x lo--' Torr, and the heater block will have risen to approximately -2o"C, with the and block are approximately -135°C and -160°C, re- heater having shut off automatically as it rose above
186
P.S. WOODS ET AL.
-75°C. Assuming that the grids are dry by this time, they can be warmed more rapidly to room temperature by moving the set-pointer on the gauge to + 25"C, with the electric-power supply still a t 5 volts. The block is not heated directly. When the temperature of the heater approaches room temperature, the power supply is turned off, the main valve is closed and the pump is turned off. The colder block and sleeveicover, a t approximately - 20°C, keep condensates from returning to the specimens. Dry N, gas is bled slowly into the chamber and when atmospheric pressure is reached, the grids are removed quickly to prevent their cooling, coated with carbon, and transferred to a desiccator containing Drierite. Care must be taken to avoid condensation of ambient moisture on the specimens.
Drying the specimens-the ethanol-substitution method (FS) This method is carried out a t atmospheric pressure without using the electric heater and sleeveicover. These unused parts can be placed outside the Dewar flask, but within the chamber to avoid disconnecting their wires. The grids are frozen in N, slush, the slush is thawed with the lamp and heat absorber a s described previously for FD, and when all the liquid N, is gone, heating is discontinued. If the grids are not in the well, they are pushed there with a small wooden stick. The block is not heated. A predetermined volume (approximately 65 ml in our apparatus) of pure anhydrous ethanol a t room temperature is poured slowly through the hole in the glass plate, using a funnel to direct it into the space between the block and the inner wall of the Dewar flask, avoiding the well where the grids are located. The absolute ethanol quickly freezes, preventing its passing through the drain-holes (Fig. 2) to the well which contains the frozen grids. However, flooding occurs automatically later when the block rises to the melting point of ethanol (- 117°C). The temperature of the alcohol and block continues to rise slowly (5.0-65°C per hour, between -117°C and -40°C) and eventually all the ice dissolves into the alcohol. Late the next day, when the temperature has reached approximately -lO°C, the heat absorber is reattached to the block and heated with the lamp until the block reaches room temperature. The grids are removed and alcohol is replaced with absolute acetone, which, in turn, is removed using a critical-point dryer and liquid CO, (Anderson, 1951). The grids are coated with carbon and stored in a desiccator.
to 0.5 mM), or dilute EDTA (0.02 to 0.10 mM). High concentrations of Mgt cause chromosomes and nuclei to condense, low concentrations cause them to expand; and causes the EDTA binds with Ca' ' and Mg' chromosomes and nuclei to expand greatly. It is difficult to control the size precisely. Root tips, each in two drops of water, or one of the above solutions, on a gelatin-subbed microscope slide, were crushed with another subbed slide (Van't Hof, 19751, and the broken cell suspension from each was collected separately in small silicone-treated watch glasses on ice. Immediately, a specially treated Formvar-filmed 200-mesh copper-rhodium index grid (cast from a 0.5% Formvarethylene dichloride solution) was introduced, film-side up, onto which the suspended material was allowed to settle for 1 hour. The grids were previously treated by glow-discharge, using 0, gas, within a Plasmod apparatus (Tegal Corp., Richmond, CAI set a t minimum power for 5 to 10 seconds, dipped in polylysine DL (O.lO%),air-dried (Mazia et al., 1975), and soaked for 5 minutes in the solution to be used for crushing the roots. Chromosomes and nuclei usually stick to the coated Formvar film, while most of the cytoplasmic debris subsequently rinses off. The grids then were passed through cold water, or one of the above dilute solutions for 15 minutes, then into cold 10% ethanol made either with water or one of the dilute solutions, and held in cold 20% ethanol (made the same way) for 30 minutes before freezing. +
+
RESULTS
Figure 3 , taken from a n earlier abstract (Woods et al., 1979), illustrates the need for using freeze-dry procedures when studying the ultrastructure of formalinfixed and expanded chromosomes. This preparation was processed by passage through increasing concentrations of acetone and dried by the CPD method (critical-point dried, meaning acetone dehydration and use of a critical-point dryer). This chromosome was expanded before dehydration by a dilute solution of EDTA (0.02 mM) while it was attached to the support film. Stereo viewing shows that only the expanded end of the chromosome is attached. Later during dehydration, the parts that were not firmly anchored became condensed; removal of water by increasing concentrations of acetone resulted in chromosome shrinkage. Observations by phase light microscopy (unpublished data) showed that 20-25% acetone is a critical concentration a t which shrinkage occurs. Under the electron microscope, condensed portions appear opaque to the Chromosomal and nuclear preparation electron beam and the internal structure cannot be The source of material was primary roots from the seen. plant Vicia faba; we describe our procedures of fixation, Figure 4 shows a chromosome that was fixed and isolation, expansion, and deposition of chromosomes expanded (0.1 mM EDTA) similarly to the one in Figand nuclei on grids. All solutions were freshly made ure 3, but this chromosome was FS from 20% ethanol; with deionized distilled water. Living seedlings were the open state is preserved. This chromosome is aptreated with 0.03% colchicine for 2 hours at 23°C to proximately 42 pm long and appears uniformly exdisrupt the spindle, then 2 mm portions of root tips panded throughout its length and width, with the rewere fixed for 10-30 minutes in ice-chilled 2% forma- sult that structures in great detail are now seen lin made with Mcleish buffer (pH 6.8); they were rinsed throughout the chromosome. Starch grains and exin water and, depending on the desired degree of ex- panded amyloplasts, partially visible here, occasionpansion (Cole, 19671, washed by passing through four 5 ally contaminate our preparations. The inset, a t the minute changes of either cold water, dilute MgC1, (0.05 same magnification, emphasizes the degree of this ex-
FREEZE-DRYING GRID-MOUNTED SPECIMENS
pansion, by showing an unexpanded metaphase chromosome that was treated with MgC1, (0.5 mM) before its attachment to the support film during FD from 20% ethanol. This chromosome is approximately 6 pm long and may represent one of the typically condensed acrocentric chromosomes of V. fuba, like the expanded one shown in the same Figure. This unexpanded chromosome is so dense that little internal detail can be seen in micrographs a t higher magnifications using 100 kV accelerating voltage. The arrows show the location of the centromere. There is no evidence of icecrystal artifact in either chromosome, even at magnifications above x 8,000. Figure 5 shows a part of a similarly enlarged fixed chromosome in stereo a t high magnification. This chromosome was expanded during its attachment to the support film and then FD from 20% ethanol. The latex spheres were deposited before the chromosome was attached. The short segments of the two chromatids, of this 42 pm-long chromosome, are sufficiently open to permit penetration of the electron beam. The internal structure of each chromatid is both fibrous and nonfibrous, the former appearing as 12-18 nm fibers while the nonfibrous ones consist of variable and much thicker, amorphous and branched material. Some of the latter appears to hold the two chromatids together. The smaller fibers are rare relative to the nonfibrous structures. There is no evidence of ice-crystal artifact in this chromosome. Figure 6 illustrates an artifact that is often seen when a chromosome is not cryoprotected with 20% ethanol but is frozen from water. However, the artifact also can occur when cryoprotected if the specimen is not blotted long enough and too much liquid is left on the grid. Parts of the chromosome are pushed aside during the formation of ice-crystals, approximately 50 nm in diameter. Removal of the crystals during drying leaves spaces. Such artifacts usually are not observed when the specimen is properly blotted from 20% ethanol (Figs. 4, 5). Figure 7 shows another artifact sometimes seen in our FD and FS preparations, in which all meaningful structure is lost. This occurs either when a chromosome dries by evaporation before freezing because of too much blotting, or dries while under vacuum, if the specimen was not completely dry during the more rapid warmup step. Surface tension apparently presses the chromosome to such an extent that macromolecular structural integrity is lost. However, gross morphology is still recognizable and the chromosome appears “ghost-like.” It is noteworthy that at the centromere (arrow), each chromatid has a distinctly denser region. Figure 8 is a low magnification micrograph of an intact nucleus from a formalin-fixed root; the specimen was treated with 0.10 mM EDTA, during attachment to the support film, and was then FS from 20% ethanol. This nucleus is approximately 50 pm in diameter. The nucleolus appears not to be expanded. The inset at the same magnification shows a normal unexpanded 10 p,m diameter nucleus treated with dilute MgC1, before FD from 20% ethanol. There is no evidence of a nuclear envelope in the expanded nucleus, although the condensed nucleus appears to have one. Because of the
187
expansion, structures become visible that are not visible in the condensed nucleus. There is no evidence at high magnifications of ice-crystal artifact in either specimen. Figure 9 is a high magnification stereo micrograph of a region near the nucleolus of the previous expanded nucleus. The 12-18 nm chromatin fibers, and the more common, much thicker nonfibrous and branched matrix material that make up most of the remaining structure of this nucleus, are clearly visible. A number of clustered 35 nm granules, and a few, very large 200 nm bodies are also visible.
DISCUSSION The ability to observe the internal structure of fixed metaphase chromosomes and nuclei satisfactorily with the conventional transmission electron microscope becomes possible when these structures are isolated, expanded, and preserved intact on grids by freeze-drying. Our results demonstrate the value of the apparatus we have constructed. It seemed that the best design would include plunging the grid with specimens attached into a cryogenic liquid such as slushy N,, within the drying chamber itself, thus avoiding any further specimen handling before complete drying. Therefore, we constructed an apparatus which is simple in design and operation, requiring little attention during the drying period. Electron microscopy of grid-mounted preparations of metaphase chromosomes has been performed with some success by many workers (DuPraw, 1970; Gall, 1966; Ris, 1976; Ris and Korenberg, 1979; Woods et al., 1977), using more conventional preparative procedures; however, in all cases, internal structure was difficult or impossible to study. Our present observations using freeze-dry methods compare favorably with these earlier findings and in addition, because the expanded state is preserved, new structures become apparent; e.g., the material holding the chromatids of metaphase chromosomes together is variable in thickness (18-65 nm), is nonfibrous, and highly branched. Uniformly thin fibers (12-18 nm), thought to be one state of chromatin, are seen within the expanded chromosomes and interphase nuclei; thicker chromatin fibers (25-30 nm) have not been identified. Also, in interphase nuclei, we observe large bodies (200 nm) and clusters of small granules (35 nm) associated with the nonfibrous and branched material. The striking feature of chromosomes and nuclei is that they are composed of such large quantities of nonfibrous and branched material, suggesting that the chromatin fibers are embedded in an amorphous matrix. Since the drying conditions in our apparatus are so different from most dryers, e.g., ones by Chiovetti et al. (1987), Coulter and Terracio (1977), Edelmann (19781, Elder et al. (1986),Linner et al. (1986), and Pawley and Ris (1987), it is useful to consider the environment under which isolated chromosomes and nuclei are dried. The 30-35°C colder condensation trap for the first 3 -4 hours of drying between -135°C and -75°C followed by a gradual warming to -20°C over the next 20 hours, at a distance of 5 mm from the specimen, a t a pressure of only 1 x Torr, are some of these conditions. The
188
P.S. WOODS ET AL.
Figs. 3-7
FREEZE-DRYING GRID-MOUNTED SPECIMENS
vacuum within the chamber of our apparatus is poor in comparison to the more sophisticated systems which use diffusion, turbo molecular, or cryogenic pumps; however, the atmosphere around the heater upon which the specimens rest is expected to be quite free from contaminants because of the cold surfaces surrounding the specimens. The ability to scavenge water and other volatile contaminants, including pump oil is quite efficient. Also the vacuum from the rotary pump is well protected from backstreaming of oils by two efficient in-line trapping devices in addition to the beaker of P,O, on the block. Under any circumstances, we judge our specimens to be clean and free of observable contaminants. Coulter and Terracio (1977), Menco (1986), and Terracio and Schwabe (1981) reviewed some theoretical aspects of drying including artifact formation. Ice-crystals usually form during initial freezing. To avoid structural damage the specimen must be cooled rapidly enough to bypass phase-changes in water during solidification. If the specimen is to be plunged into a cryogenic liquid such as slushy N2, it is important that the total mass of the specimen, including its support, be a s small as possible, and that it makes good thermal contact with the cryogen as quickly as possible. Grids with very small and thin subcellular structures attached,
Fig. 3. Stereo micrograph showing the shrinking effect of acetone dehydration on a n EDTA-expanded, formalin-fixed, V. faba metaphase chromosome processed by the CPD method. This chromosome was expanded and attached to the support film before dehydration, and as we see here, after dehydration, only the part firmly anchored is maintained in the expanded state. Total tilt 12". Bar, 1.0 &m. x 12,800. Fig. 4. Low magnification micrograph of a n expanded, metaphase chromosome, treated with EDTA to increase its size approximately seven-fold before attachment to the support film, and then FS from 20% ethanol. Here we see preservation of the expanded state for the entire chromosome. Inset: At the same magnification, inset emphasizes the degree of this expansion, by showing an unexpanded, typically condensed metaphase chromosome that was treated with dilute MgCI, and FD from 20% ethanol. Arrows indicate the location of the centromere. Bar for both chromosomes, 5.0 bm. x 3,000. Fig. 5. High magnification stereo micrograph showing part of a seven-fold expanded metaphase chromosome treated with EDTA a t the time of attachment to the support film and FD from 20R, ethanol. Here we see that the internal structure of each chromatid is composed mainly of 18-65 nm nonfibrous and branched material. The 0.36 pm diameter latex spheres were deposited before the chromosome was attached. Total tilt 16". Bar, 1.0 pm. x 16,900. Fig. 6. Micrograph of an unexpanded, typically condensed metaphase chromosome, treated similarly to the one shown in the inset of Figure 4,but FD from water, instead of 20% ethanol. Here we see ice-crystal artifacts throughout the chromosome. Bar, 1.O pm. x 9,400. Fig. 7. Micrograph of an unexpanded, typically condensed metaphase chromosome, treated similarly to the one in the inset of Figure 4,but purposefully air-dried from 20% ethanol before freezing, and then FD. Here we see complete loss of fine structure caused by surface tension forces during evaporative-drying; however, gross morphology is still recognizable. The obvious increase in size is explained by spreading of the mass, apparently without loss, over a larger area. The arrow indicates the location of the centromere. Bar, 1.0 pm. x 4,100.
189
approach the ideal when frozen by propulsion into the super-cooled liquid phase (-2lO'C) of slushy N,, without the formation of a n insulating gas layer (Leidenfrost effect) over the surface of the specimen. Ice-crystals also can form in a n apparently well frozen specimen during drying after freezing and partial warming, beginning a t about - 130°C. This process is called devitrification, where solid water becomes a supercooled liquid and small crystals begin to form. Apparently, this process proceeds slowly, and by -90"C, any resulting artifact a s seen in the electron microscope may still not be serious. Large crystals that are so objectionable because of the artifacts they produce, apparently form a t temperatures above -9O"C, during a process called recrystallization, where unstable small crystals grow to a more stable larger size. Another important consideration is the cooling effect on the specimen d x i n g drying as water molecules sublime. The surface of the specimen is colder than the support upon which it rests. The rate of sublimation affects the temperature of the specimen; the higher the rate, the greater the cooling. This cooling effect then slows the rate of sublimation, until the rate comes to equilibrium for a given temperature. Vapor pressure also influences the rate of sublimation and must not become a limiting factor. The evacuating system, including the water-vapor trap, must maintain a n unsaturated vapor pressure. From a n optimal drying curve for ice in a high vacuum (supplied by Denton Vacuum, Inc.), the rate of sublimation rises rapidly as the equilibrium temperature rises. For example, a t -130°C the rate as calculated from their curve is negligible a t 0.01 pm per day; at -108"C, it is about 0.6 pm per hour, and at -75°C the rate is about 600 p.m per hour. Some dryers, e.g., the one designed by Pawley and Ris (1987) are used at -123°C for 2 days, a t high vacuum, followed by a gradual 1-day warmup to room temperature. These workers show micrographs a t high magnification of grid-mounted cells in culture completely free of artifact when the specimen is cushion slam-frozen against a liquid He-chilled bar. In their case, drying may occur mainly during the early stages of warming. Coulter and Terracio (1977) present evidence from partial pressure observations that sublimation is complete within 5 hours when blocks of tissue rise stepwise between -130°C and -80°C. More recently, Linner et al. (1986)using a special partial pressure analyzer at a much higher vacuum, showed that sublimation was complete within 7 hours, when the temperature was raised from -128°C to -121"C, temperatures where devitrification is thought not to occur. Both groups of workers showed excellent micrographs of thin sectioned material that was frozen against a liquid N, chilled metal surface. Other data by Denton Vacuum, Inc. show at the other extreme, e.g., a t -6O"C, that the rate of sublimation of ice under optimal conditions is approximately 5,000 pm per hour o r about 80 pm per minute. If the specimen is not torn apart, it would be dry before recrystallization could become a problem. Obviously, drying of tissue after freezing does not occur suddenly. The specimen dries gradually as i t is warmed from the supercold freezing temperature, and would be dry by the time it reached
190
P.S. WOODS ET AL.
Fig. 8. Low magnification micrograph of an expanded nucleus treated with EDTA to increase its diameter five-fold before attachment to the support film and then FS from 20%' ethanol. Inset: At the same magnification, inset emphasizes the degree of this expansion, by showing an unexpanded nucleus treated with dilute MgC1, before FD from 20% ethanol. Shown are: nucleolus (Nu), chromatin (Ch), and contaminating cytoplasmic starch grains (3.Bar for both nuclei, 5 Fm. ~ 1 , 4 0 0 .
Fig. 9. High magnification stereo micrograph of a part of the expanded nucleus of the previous figure. The 18-65 nm nonfibrous and branched matrix material (MI, and the less common 12-18 nm chromatin fibers (F) make up most of the fine structure in this field of view. Also present are 35 nm granules ( G )in clusters, and large 200 nm nuclear bodies (Nb). Total tilt 16". Bar, 1.0 km. x 20,600.
the warmer temperature when structure might otherwise be destroyed. We often have successfully dried other kinds of grid-mounted specimens when the heater was set at -60°C; e.g., with fixed cells in culture, we observed very little damage or abnormal structure within cells near their thinner margins when they were frozen, warmed, and held for a time a s we
describe, at this higher temperature. This observation shows that these thinner parts were already dry before they reached the warmer temperature. The degree to which our apparatus satisfies all of the requirements thought to be necessary for proper drying is uncertain; however, a majority of our preparations are free of observable artifacts.
FREEZE-DRYING GRID-MOUNTED SPECIMENS
ACKNOWLEDGMENTS This work was supported in part by the U S . Department of Energy, Office of Health and Environmental Research, and by a grant to the senior author from the City University of New York PSC-CUNY Research Award Program, when he was a member of the Biology Department, Queens College, CUNY. Funds allocated to Jack Van’t Hof of the Biology Department, Brookhaven National Laboratory also contributed to the construction of the apparatus. Testing of prototypes began in 1980 when the senior author was on sabbatical in the Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder. The support of Mircea Fotino, the NIH high voltage EM users grant, and the assistance of George Wray was greatly appreciated. The authors also would like to thank Avril Woodhead for her suggestions during writing of the manuscript. REFERENCES Anderson, T. (1951) Techniques for the preservation of three-dimensional structures in preparing specimens for the electron microscope. Trans. N. Y. Acad. Sci. Ser. 11, 13:130-134. Chiovetti, R., McGuffee, L.J., Little, S.A., Wheeler-Clark, E., and Brass-Dale, J . (1987) Combined quick freezing, freeze-drying, and embedding tissue a t low temperatures and in low viscosity resins. J . Elec. Microsc. Tech., 5:l-15. Cole, A. (1967) Chromosome structure. In: Theoretical and Experimental Biophysics. Vol. 1. A. Cole, ed. Marcel Dekker, New York, pp. 305-375. Coulter, H.D., and Terracio, L. (1977) Preparation of biological tissues for electron microscopy by freeze-drying. Anat. Rec., 187:477-494. Coulter, H.D., and Terracio, L. (1978) An all glass freeze-dryer for TEM specimens with a n improved design for temperature regulation, fixation, and infiltration. 9th International Congress on Electron Microscopy, 11:60-61. DuPraw, E.J. (1970) DNA and Chromosomes. Holt, Rinehart and Winston, Inc., New York, 340 pp. Edelman, L. (1978) A simple freeze-drying technique for preparing
191
biological tissue without chemical fixation for electron microscopy. J . Microsc., 112:243-248. Elder, H.Y., Biddlecombe, W.H., Tetley, L., Wilson, S.M., and McEwan Jenkinson, D. (1986) Construction of low temperature freeze driers. EMSA Bull., 16:lll-113. Gall, J.G. (1966) Chromosome fibers by a spreading technique. Chromosoma, 20:221-223. Linner, J.G., Livesey, S.A., Harrison, D.S., and Steiner, A.L. (1986) A new technique for removal of amorphous phase tissue water without ice crystal damage: a preoperative method for ultrastructural analysis and immunoelectron microscopy. J. Histochem. Cytochem., 34:1123-1135. Mazia, D., Schatten, G., and Winfield, S. (1975) Adhesion of cells to surfaces coated with polylysine. J . Cell. Biol., 66:198-200. Menco, B., Ph. M. (1986) A survey of ultra-rapid cryofixation methods with particular emphasis on applications to freeze-fracturing, freeze-etching, and freeze-substitution. J. Elec. Microsc. Tech., 4: 177-240. Pawley, J . , and Ris, H. (1987) Structure of the cytoplasmic filament system in freeze-dried whole mounts viewed by HVEM. J. Microsc., 145:319-332. Phillips, T.E., and Boyne, A.F. (1984) Liquid nitrogen-based quick freezing: Experiences with bounce-free delivery of cholinergic nerve terminals to a metal surface. J . Elec. Microsc. Tech., 1:9-29. Ris, H. (1976) Levels of chromosome organization, Electron Micros. Proceedings of the 6th European Congress on Electron Microscopy, 1976, Vol. 11, pp. 21-25. Ris, H., Korenberg, J. (1979) Chromosome structure and levels of chromosome organization. In: Cell Biology A Comprehensive Treatise. Vol. 2. The Structure and Replication of Genetic Material. D.M. Prescott and L. Goldstein, eds. Academic Press, New York, pp. 267-361. Terracio, L., and Schwabe, K.G. (1981) Freezing and drying of biological tissues for electron microscopy. J. Histochem. Cytochem., 29: 1021-1028. Van’t Hof, J . (1975) DNA Fiber replication in chromosomes of a higher plant (Pzsurn satiuurn). Exp. Cell Res., 93:95-104. Woods, P.S., Ledbetter, M.C., and Van’t Hof, J . (1979) Use of EDTA in studies of isolated metaphase chromosome ultrastructure: visualization of a nonfibrous component. Proceedings of the 37th Annual EMSA Meeting, pp. 36-67. Woods, P.S., Van’t Hof, J., and Ledbetter, M.C. (1977) A differential centrifugation procedure for obtaining unstretched metaphase chromosomes from higher plants for transmission electron microscopy. Proceedings of the 35th Annual EMSA Meeting, pp. 402-403.
Preservation of EDTA-Expanded Grid-Mounted Chromosomes and Nuclei for Electron Microscopy Using a Specially Designed Freeze-Dryer PHILIP S. WOODS, MYRON C. LEDBETTER, AND NEAL TEMPEL Biology Department, Brookhaven National Laboratory, Upton, New York 11973
KEY WORDS
Whole-mounts, Freeze-dry, Freeze-substitution, TEM
ABSTRACT We describe methods for freezing and drying EDTA-expanded, fixed metaphase chromosomes and nuclei, attached to grids as whole-mounts, for transmission electron microscopy. These methods use a special apparatus that is simple to construct. While separate freezers and dryers are commercially available, one for freezing blocks of tissue by slamming them against a cold metal surface, and the other for vacuum drying the frozen tissue, our apparatus is designed for gentler, cryogenic liquid plunge freezing and drying, sequentially, in the same apparatus, thus avoiding any compression or damage to the sepcimen. Use of a cryoprotectant is not essential; however, good results are obtained more often when 20% ethanol is used. Freezing is accomplished by rapid propulsion of the grid, with specimens attached, into slushy N, (-210°C) within the drying chamber; drying is automatic, by either sublimation under vacuum or by solvent substitution using absolute ethanol followed by acetone, which, in turn, is removed with a critical-point dryer. The apparatus offers a means of drying chromosomes and nuclei in a n expanded state, and avoids the shrinkage of these structures that occurs during stepwise passage through increasing concentrations of ethanol or acetone. INTRODUCTION
MATERIALS AND METHODS The apparatus When isolated formalin-fixed metaphase chromoFigure 1 shows our apparatus and the inset shows a somes are examined as whole-mounts by the 100 kv disassembled view of the electric heater, and the transmission electron microscope, they are nearly sleeve/cover assembly, the latter serving as a wateropaque to the electron beam (Woods et al., 1977) and vapor condensing trap located in the block. Figure 2 is reveal little of their internal structure. We have at- a sectional in-scale diagram outlining the major comtempted to render these chromosomes more accessible ponents of the apparatus within the vacuum chamber. to electron microscopy by expanding them with the This design uses a modified 9-port freeze-etch collar chelating agent Na,EDTA.2H20 (Woods et al., 1979). from the D F E S apparatus (Denton Vacuum, Inc., This treatment reduced their density; however, subse- Cherry Hill, NJ). We mount the collar separately on a n quent dehydration with acetone or ethanol always aluminum base plate. The collar is evacuated with a caused the chromosomes to recondense again. Clearly, rotary vacuum pump using a liquid Nz trap to protect another method of drying was needed to preserve the the specimen from backstreaming of oil. With this modchromosomes in a n open state. Freeze-drying was con- est apparatus we have obtained usable specimens a templated. Several dryers had been either described or majority of the time. were available commercially, e.g., ones by Chiovetti et al. (1987)) Coulter and Terracio (19781, Edelmann Freezing the specimen (1978)) Elder et al. (1986), and Linner e t al. (1986); however, these seemed inappropriate for grid-mounted 1. The liquid cryogen used is slushy N2 (-210°C) specimens, since they usually were used with “slammade by first filling with liquid N, the Dewar flask (I), frozen” blocks of tissue for eventual embedding and sectioning (Phillips and Boyne, 1984). Pawley and Ris (1987) used a commercial slamming device which they modified with a cushion for freezing cells growing on grids; the grids then were transferred to a dryer of their Received October 4, 1 9 8 9 accepted in revised form August 25, 1990. design. Our apparatus is simple to construct, and proAddress reprint requests to Dr. Philip S. Woods, Biology Department, vides for gentle “plunge-freezing” and subsequent dry- Brookhaven National Laboratory, Upton, NY 11973. ing in the same apparatus, avoiding compression of the Philip S.Woods is a guest scientist of the Biology Department, Brookhaven specimen and damage to it during transfer. This paper National Laboratory, Upton, NY 11973. Abbreviations: FD = freeze-dried, meaning frozen from either water or 20% describes the apparatus and its use, especially in the ethanol and dried by vacuum sublimation; FS = freeze-substitution, meaning study of EDTA-expanded, fixed chromosomes and nu- frozen from either water or 20% ethanol and dried by ethanol substitution and use of a critical-point dryer and acetone. clei isolated from growing root tips of a higher plant.
PUBLISHED 1991 WILEY-LISS, INC.
184
P.S. WOODS ET AL.
Fig. 1. The freeze-dry apparatus consisting of: (A) Denton freezeetch 9-port collar, (BJ aluminum base plate, (C) optical meter relay, double set-point, thermocouple, temperature gauge, (D) direct current power supply, (EJ dry-ice chilled Drierite water-vapor trap, (F) inlet hose from high pressure N, tank, (GI inlet valve to chamber for dry N, gas, (H) main valve to chamber from mechanical pump, (I) Dewar flask, (J)heat absorber with copper post, (K) movable arm for lifting and lowering heat absorber and post, (L) block thermocouple wire, (M) heater thermocouple wire, (N) Copper-Constantan thermocouple
plugs, (0)plastic beaker of P,O,, (P) Lucite cover to port for electric heater and thermocouple wires. Inset: Shows the electric heatersleeveicover assembly consisting of: (A) the heater (its plastic screen fence is removed), two insulated stainless steel electrical lead wires, and the heater thermocouple wire, that pass through holes in the top of the sleeveicover (B).When the heater is enclosed within the sleeve/ cover, and the two, in turn, are fitted into the large well of the block (as shown in Fig. 21, the heater does not touch the sleevdcover nor does it touch the block except by the needle legs.
containing the 3 kg copper block and heater-sleeve/ cover assembly. After nucleate boiling subsides, the chamber is evacuated with a rotary pump (112 hp, Duo Seal, 160 liters per minute capacity) until the nitrogen is almost completely frozen. The temperature of the block should read -217"C, about 2" warmer than the lowest temperature attainable. The chamber then is brought to atmospheric pressure within 20 seconds with dry N, gas from a high pressure cylinder (not shown) equipped with a pressure reducer, using the small inlet valve (G) on the collar. This gas is cooled and dried a s it passes through the dry-ice chilled Drierite (anhydrous CaSO,) water-vapor trap (E). 2. Grids are picked up with forceps from ice-chilled 20% ethanol and placed, specimen side up, on the end of a long thin wooden stick, blotted briefly at one edge, and together turned and snapped firmly by bending and releasing the end of the stick over the hole in the glass plate. The grid is propelled a t high velocity into the slush. It is important that freezing of the grids
begins within the first few seconds after atmospheric pressure is reached before the temperature a t the surface of the slush becomes too high. A plastic screen fence (Fig. 21, wrapped around the block, prevents the grids from falling under the block. A CopperConstantan thermocouple is mounted in the block to monitor the temperature of the block with its wire passing through a port in the collar connecting by way of wire (L) and a set of Copper-Constantan thermocouple plugs (N) to a double (or single) set- point, thermocouple, temperature gauge (C) made by A.P.I. Instrument Co., Clinton, MA. 3. The remaining liquid N, is boiled away by raising the block temperature. This is done by attaching the post of the heat absorber (J)to the heat conductor on the block (Fig. 2) using the movable arm (K) and heating the absorber with a heat lamp (not shown). A piece of lead is fused to the corner (underside) of the heat absorber opposite its post (Fig. 2) to serve as a counterbalance.
FREEZE-DRYING GRID-MOUNTED SPECIMENS
185
Fig. 2. Diagrammatic view of the part of the apparatus in the vacuum chamber.
Drying the specimens-the vacuum sublimation method (FD)
spectively, the electric heater is turned on to dry the specimens. This is accomplished as follows: the switch on the direct-current power supply (2 amps, 10 volts 1. Grids are pushed through the holes of the sleeve1 capacity) is turned on. Current to the embedded resiscover (Fig. 1, inset and 2 ) where they fall directly onto tor in the heater (Fig. 1, inset) is regulated by the setthe electric heater, which consists of a 30 ohm, 5 watt point temperature gauge. For this the heater thermoresistor embedded in a small copper cylinder with sil- couple is connected to the gauge and the set-pointer on icone rubber adhesive, and thermally isolated from the the gauge is moved to a suitable drying temperature. block with three needle-sharp stainless steel legs. An- We find that -75°C yields good results. The amount of other smaller plastic screen (not shown) is wrapped heat is controlled by the voltage of the power supply, around the heater to prevent grids from falling under and the set-pointer determines the temperature a t it. There is another Copper-Constantan thermocouple which the heater will be maintained. We find that 5 mounted in the heater with its wire passing through volts raises the temperature of the heater from - 135°C another port on the collar (P) to the wire (M) that may to -75°C in about 2 hours. The temperature of the be connected to the same temperature gauge (C) by block and sleeveicover rises a t a much slower rate and another Copper-Constantan thermocouple plug (N). both are always colder than the heater. The set-pointer Exchange of these plugs between the two thermocou- gauge maintains the heater constantly a t the chosen ples allows the temperature of the block or heater to be temperature of -75°C until the block later approaches read by the same temperature gauge, provided the to- -75"C, at which time the heater automatically shuts tal lengths of each thermocouple wire are equal. After off. During drying the colder sleeveicover traps water the grids are placed on the heater, a small beaker of the molecules as they sublime from the grids and the specdesiccant P,O, (0)is passed through the hole in the imens gradually dry. The cycling time is 1 to 2 hours glass plate and placed on the block. When all the liquid when the temperature is held at -75°C. Denton VacN, is gone, the heat lamp is turned off and the heat uum, Inc. informs us (personal communication) that absorber with post is disconnected. the rate of sublimation of ice is approximately 10 pm 2 . The hole in the glass plate is covered with a plug per minute at -75"C, under optimum conditions of vacand the chamber is evacuated. The pump is left on uum and vapor pressure. Since our specimens are less during the entire drying period. In addition to the liq- than 100 pm in thickness (including the ice over the uid N2 trap on the vacuum line to the pump, there is an expanded nuclei), they are dry within the first 10 minoptic-dense foreline trap made by Ion Equipment Corp., utes if these data hold for the conditions in our appaSanta Clara, CA (both traps not shown). ratus. 3. After 5-10 minutes, when the vacuum comes 4. By the afternoon of the next day the heater and close t o its best value of 1 x lo--' Torr, and the heater block will have risen to approximately -2o"C, with the and block are approximately -135°C and -160°C, re- heater having shut off automatically as it rose above
186
P.S. WOODS ET AL.
-75°C. Assuming that the grids are dry by this time, they can be warmed more rapidly to room temperature by moving the set-pointer on the gauge to + 25"C, with the electric-power supply still a t 5 volts. The block is not heated directly. When the temperature of the heater approaches room temperature, the power supply is turned off, the main valve is closed and the pump is turned off. The colder block and sleeveicover, a t approximately - 20°C, keep condensates from returning to the specimens. Dry N, gas is bled slowly into the chamber and when atmospheric pressure is reached, the grids are removed quickly to prevent their cooling, coated with carbon, and transferred to a desiccator containing Drierite. Care must be taken to avoid condensation of ambient moisture on the specimens.
Drying the specimens-the ethanol-substitution method (FS) This method is carried out a t atmospheric pressure without using the electric heater and sleeveicover. These unused parts can be placed outside the Dewar flask, but within the chamber to avoid disconnecting their wires. The grids are frozen in N, slush, the slush is thawed with the lamp and heat absorber a s described previously for FD, and when all the liquid N, is gone, heating is discontinued. If the grids are not in the well, they are pushed there with a small wooden stick. The block is not heated. A predetermined volume (approximately 65 ml in our apparatus) of pure anhydrous ethanol a t room temperature is poured slowly through the hole in the glass plate, using a funnel to direct it into the space between the block and the inner wall of the Dewar flask, avoiding the well where the grids are located. The absolute ethanol quickly freezes, preventing its passing through the drain-holes (Fig. 2) to the well which contains the frozen grids. However, flooding occurs automatically later when the block rises to the melting point of ethanol (- 117°C). The temperature of the alcohol and block continues to rise slowly (5.0-65°C per hour, between -117°C and -40°C) and eventually all the ice dissolves into the alcohol. Late the next day, when the temperature has reached approximately -lO°C, the heat absorber is reattached to the block and heated with the lamp until the block reaches room temperature. The grids are removed and alcohol is replaced with absolute acetone, which, in turn, is removed using a critical-point dryer and liquid CO, (Anderson, 1951). The grids are coated with carbon and stored in a desiccator.
to 0.5 mM), or dilute EDTA (0.02 to 0.10 mM). High concentrations of Mgt cause chromosomes and nuclei to condense, low concentrations cause them to expand; and causes the EDTA binds with Ca' ' and Mg' chromosomes and nuclei to expand greatly. It is difficult to control the size precisely. Root tips, each in two drops of water, or one of the above solutions, on a gelatin-subbed microscope slide, were crushed with another subbed slide (Van't Hof, 19751, and the broken cell suspension from each was collected separately in small silicone-treated watch glasses on ice. Immediately, a specially treated Formvar-filmed 200-mesh copper-rhodium index grid (cast from a 0.5% Formvarethylene dichloride solution) was introduced, film-side up, onto which the suspended material was allowed to settle for 1 hour. The grids were previously treated by glow-discharge, using 0, gas, within a Plasmod apparatus (Tegal Corp., Richmond, CAI set a t minimum power for 5 to 10 seconds, dipped in polylysine DL (O.lO%),air-dried (Mazia et al., 1975), and soaked for 5 minutes in the solution to be used for crushing the roots. Chromosomes and nuclei usually stick to the coated Formvar film, while most of the cytoplasmic debris subsequently rinses off. The grids then were passed through cold water, or one of the above dilute solutions for 15 minutes, then into cold 10% ethanol made either with water or one of the dilute solutions, and held in cold 20% ethanol (made the same way) for 30 minutes before freezing. +
+
RESULTS
Figure 3 , taken from a n earlier abstract (Woods et al., 1979), illustrates the need for using freeze-dry procedures when studying the ultrastructure of formalinfixed and expanded chromosomes. This preparation was processed by passage through increasing concentrations of acetone and dried by the CPD method (critical-point dried, meaning acetone dehydration and use of a critical-point dryer). This chromosome was expanded before dehydration by a dilute solution of EDTA (0.02 mM) while it was attached to the support film. Stereo viewing shows that only the expanded end of the chromosome is attached. Later during dehydration, the parts that were not firmly anchored became condensed; removal of water by increasing concentrations of acetone resulted in chromosome shrinkage. Observations by phase light microscopy (unpublished data) showed that 20-25% acetone is a critical concentration a t which shrinkage occurs. Under the electron microscope, condensed portions appear opaque to the Chromosomal and nuclear preparation electron beam and the internal structure cannot be The source of material was primary roots from the seen. plant Vicia faba; we describe our procedures of fixation, Figure 4 shows a chromosome that was fixed and isolation, expansion, and deposition of chromosomes expanded (0.1 mM EDTA) similarly to the one in Figand nuclei on grids. All solutions were freshly made ure 3, but this chromosome was FS from 20% ethanol; with deionized distilled water. Living seedlings were the open state is preserved. This chromosome is aptreated with 0.03% colchicine for 2 hours at 23°C to proximately 42 pm long and appears uniformly exdisrupt the spindle, then 2 mm portions of root tips panded throughout its length and width, with the rewere fixed for 10-30 minutes in ice-chilled 2% forma- sult that structures in great detail are now seen lin made with Mcleish buffer (pH 6.8); they were rinsed throughout the chromosome. Starch grains and exin water and, depending on the desired degree of ex- panded amyloplasts, partially visible here, occasionpansion (Cole, 19671, washed by passing through four 5 ally contaminate our preparations. The inset, a t the minute changes of either cold water, dilute MgC1, (0.05 same magnification, emphasizes the degree of this ex-
FREEZE-DRYING GRID-MOUNTED SPECIMENS
pansion, by showing an unexpanded metaphase chromosome that was treated with MgC1, (0.5 mM) before its attachment to the support film during FD from 20% ethanol. This chromosome is approximately 6 pm long and may represent one of the typically condensed acrocentric chromosomes of V. fuba, like the expanded one shown in the same Figure. This unexpanded chromosome is so dense that little internal detail can be seen in micrographs a t higher magnifications using 100 kV accelerating voltage. The arrows show the location of the centromere. There is no evidence of icecrystal artifact in either chromosome, even at magnifications above x 8,000. Figure 5 shows a part of a similarly enlarged fixed chromosome in stereo a t high magnification. This chromosome was expanded during its attachment to the support film and then FD from 20% ethanol. The latex spheres were deposited before the chromosome was attached. The short segments of the two chromatids, of this 42 pm-long chromosome, are sufficiently open to permit penetration of the electron beam. The internal structure of each chromatid is both fibrous and nonfibrous, the former appearing as 12-18 nm fibers while the nonfibrous ones consist of variable and much thicker, amorphous and branched material. Some of the latter appears to hold the two chromatids together. The smaller fibers are rare relative to the nonfibrous structures. There is no evidence of ice-crystal artifact in this chromosome. Figure 6 illustrates an artifact that is often seen when a chromosome is not cryoprotected with 20% ethanol but is frozen from water. However, the artifact also can occur when cryoprotected if the specimen is not blotted long enough and too much liquid is left on the grid. Parts of the chromosome are pushed aside during the formation of ice-crystals, approximately 50 nm in diameter. Removal of the crystals during drying leaves spaces. Such artifacts usually are not observed when the specimen is properly blotted from 20% ethanol (Figs. 4, 5). Figure 7 shows another artifact sometimes seen in our FD and FS preparations, in which all meaningful structure is lost. This occurs either when a chromosome dries by evaporation before freezing because of too much blotting, or dries while under vacuum, if the specimen was not completely dry during the more rapid warmup step. Surface tension apparently presses the chromosome to such an extent that macromolecular structural integrity is lost. However, gross morphology is still recognizable and the chromosome appears “ghost-like.” It is noteworthy that at the centromere (arrow), each chromatid has a distinctly denser region. Figure 8 is a low magnification micrograph of an intact nucleus from a formalin-fixed root; the specimen was treated with 0.10 mM EDTA, during attachment to the support film, and was then FS from 20% ethanol. This nucleus is approximately 50 pm in diameter. The nucleolus appears not to be expanded. The inset at the same magnification shows a normal unexpanded 10 p,m diameter nucleus treated with dilute MgC1, before FD from 20% ethanol. There is no evidence of a nuclear envelope in the expanded nucleus, although the condensed nucleus appears to have one. Because of the
187
expansion, structures become visible that are not visible in the condensed nucleus. There is no evidence at high magnifications of ice-crystal artifact in either specimen. Figure 9 is a high magnification stereo micrograph of a region near the nucleolus of the previous expanded nucleus. The 12-18 nm chromatin fibers, and the more common, much thicker nonfibrous and branched matrix material that make up most of the remaining structure of this nucleus, are clearly visible. A number of clustered 35 nm granules, and a few, very large 200 nm bodies are also visible.
DISCUSSION The ability to observe the internal structure of fixed metaphase chromosomes and nuclei satisfactorily with the conventional transmission electron microscope becomes possible when these structures are isolated, expanded, and preserved intact on grids by freeze-drying. Our results demonstrate the value of the apparatus we have constructed. It seemed that the best design would include plunging the grid with specimens attached into a cryogenic liquid such as slushy N,, within the drying chamber itself, thus avoiding any further specimen handling before complete drying. Therefore, we constructed an apparatus which is simple in design and operation, requiring little attention during the drying period. Electron microscopy of grid-mounted preparations of metaphase chromosomes has been performed with some success by many workers (DuPraw, 1970; Gall, 1966; Ris, 1976; Ris and Korenberg, 1979; Woods et al., 1977), using more conventional preparative procedures; however, in all cases, internal structure was difficult or impossible to study. Our present observations using freeze-dry methods compare favorably with these earlier findings and in addition, because the expanded state is preserved, new structures become apparent; e.g., the material holding the chromatids of metaphase chromosomes together is variable in thickness (18-65 nm), is nonfibrous, and highly branched. Uniformly thin fibers (12-18 nm), thought to be one state of chromatin, are seen within the expanded chromosomes and interphase nuclei; thicker chromatin fibers (25-30 nm) have not been identified. Also, in interphase nuclei, we observe large bodies (200 nm) and clusters of small granules (35 nm) associated with the nonfibrous and branched material. The striking feature of chromosomes and nuclei is that they are composed of such large quantities of nonfibrous and branched material, suggesting that the chromatin fibers are embedded in an amorphous matrix. Since the drying conditions in our apparatus are so different from most dryers, e.g., ones by Chiovetti et al. (1987), Coulter and Terracio (1977), Edelmann (19781, Elder et al. (1986),Linner et al. (1986), and Pawley and Ris (1987), it is useful to consider the environment under which isolated chromosomes and nuclei are dried. The 30-35°C colder condensation trap for the first 3 -4 hours of drying between -135°C and -75°C followed by a gradual warming to -20°C over the next 20 hours, at a distance of 5 mm from the specimen, a t a pressure of only 1 x Torr, are some of these conditions. The
188
P.S. WOODS ET AL.
Figs. 3-7
FREEZE-DRYING GRID-MOUNTED SPECIMENS
vacuum within the chamber of our apparatus is poor in comparison to the more sophisticated systems which use diffusion, turbo molecular, or cryogenic pumps; however, the atmosphere around the heater upon which the specimens rest is expected to be quite free from contaminants because of the cold surfaces surrounding the specimens. The ability to scavenge water and other volatile contaminants, including pump oil is quite efficient. Also the vacuum from the rotary pump is well protected from backstreaming of oils by two efficient in-line trapping devices in addition to the beaker of P,O, on the block. Under any circumstances, we judge our specimens to be clean and free of observable contaminants. Coulter and Terracio (1977), Menco (1986), and Terracio and Schwabe (1981) reviewed some theoretical aspects of drying including artifact formation. Ice-crystals usually form during initial freezing. To avoid structural damage the specimen must be cooled rapidly enough to bypass phase-changes in water during solidification. If the specimen is to be plunged into a cryogenic liquid such as slushy N2, it is important that the total mass of the specimen, including its support, be a s small as possible, and that it makes good thermal contact with the cryogen as quickly as possible. Grids with very small and thin subcellular structures attached,
Fig. 3. Stereo micrograph showing the shrinking effect of acetone dehydration on a n EDTA-expanded, formalin-fixed, V. faba metaphase chromosome processed by the CPD method. This chromosome was expanded and attached to the support film before dehydration, and as we see here, after dehydration, only the part firmly anchored is maintained in the expanded state. Total tilt 12". Bar, 1.0 &m. x 12,800. Fig. 4. Low magnification micrograph of a n expanded, metaphase chromosome, treated with EDTA to increase its size approximately seven-fold before attachment to the support film, and then FS from 20% ethanol. Here we see preservation of the expanded state for the entire chromosome. Inset: At the same magnification, inset emphasizes the degree of this expansion, by showing an unexpanded, typically condensed metaphase chromosome that was treated with dilute MgCI, and FD from 20% ethanol. Arrows indicate the location of the centromere. Bar for both chromosomes, 5.0 bm. x 3,000. Fig. 5. High magnification stereo micrograph showing part of a seven-fold expanded metaphase chromosome treated with EDTA a t the time of attachment to the support film and FD from 20R, ethanol. Here we see that the internal structure of each chromatid is composed mainly of 18-65 nm nonfibrous and branched material. The 0.36 pm diameter latex spheres were deposited before the chromosome was attached. Total tilt 16". Bar, 1.0 pm. x 16,900. Fig. 6. Micrograph of an unexpanded, typically condensed metaphase chromosome, treated similarly to the one shown in the inset of Figure 4,but FD from water, instead of 20% ethanol. Here we see ice-crystal artifacts throughout the chromosome. Bar, 1.O pm. x 9,400. Fig. 7. Micrograph of an unexpanded, typically condensed metaphase chromosome, treated similarly to the one in the inset of Figure 4,but purposefully air-dried from 20% ethanol before freezing, and then FD. Here we see complete loss of fine structure caused by surface tension forces during evaporative-drying; however, gross morphology is still recognizable. The obvious increase in size is explained by spreading of the mass, apparently without loss, over a larger area. The arrow indicates the location of the centromere. Bar, 1.0 pm. x 4,100.
189
approach the ideal when frozen by propulsion into the super-cooled liquid phase (-2lO'C) of slushy N,, without the formation of a n insulating gas layer (Leidenfrost effect) over the surface of the specimen. Ice-crystals also can form in a n apparently well frozen specimen during drying after freezing and partial warming, beginning a t about - 130°C. This process is called devitrification, where solid water becomes a supercooled liquid and small crystals begin to form. Apparently, this process proceeds slowly, and by -90"C, any resulting artifact a s seen in the electron microscope may still not be serious. Large crystals that are so objectionable because of the artifacts they produce, apparently form a t temperatures above -9O"C, during a process called recrystallization, where unstable small crystals grow to a more stable larger size. Another important consideration is the cooling effect on the specimen d x i n g drying as water molecules sublime. The surface of the specimen is colder than the support upon which it rests. The rate of sublimation affects the temperature of the specimen; the higher the rate, the greater the cooling. This cooling effect then slows the rate of sublimation, until the rate comes to equilibrium for a given temperature. Vapor pressure also influences the rate of sublimation and must not become a limiting factor. The evacuating system, including the water-vapor trap, must maintain a n unsaturated vapor pressure. From a n optimal drying curve for ice in a high vacuum (supplied by Denton Vacuum, Inc.), the rate of sublimation rises rapidly as the equilibrium temperature rises. For example, a t -130°C the rate as calculated from their curve is negligible a t 0.01 pm per day; at -108"C, it is about 0.6 pm per hour, and at -75°C the rate is about 600 p.m per hour. Some dryers, e.g., the one designed by Pawley and Ris (1987) are used at -123°C for 2 days, a t high vacuum, followed by a gradual 1-day warmup to room temperature. These workers show micrographs a t high magnification of grid-mounted cells in culture completely free of artifact when the specimen is cushion slam-frozen against a liquid He-chilled bar. In their case, drying may occur mainly during the early stages of warming. Coulter and Terracio (1977) present evidence from partial pressure observations that sublimation is complete within 5 hours when blocks of tissue rise stepwise between -130°C and -80°C. More recently, Linner et al. (1986)using a special partial pressure analyzer at a much higher vacuum, showed that sublimation was complete within 7 hours, when the temperature was raised from -128°C to -121"C, temperatures where devitrification is thought not to occur. Both groups of workers showed excellent micrographs of thin sectioned material that was frozen against a liquid N, chilled metal surface. Other data by Denton Vacuum, Inc. show at the other extreme, e.g., a t -6O"C, that the rate of sublimation of ice under optimal conditions is approximately 5,000 pm per hour o r about 80 pm per minute. If the specimen is not torn apart, it would be dry before recrystallization could become a problem. Obviously, drying of tissue after freezing does not occur suddenly. The specimen dries gradually as i t is warmed from the supercold freezing temperature, and would be dry by the time it reached
190
P.S. WOODS ET AL.
Fig. 8. Low magnification micrograph of an expanded nucleus treated with EDTA to increase its diameter five-fold before attachment to the support film and then FS from 20%' ethanol. Inset: At the same magnification, inset emphasizes the degree of this expansion, by showing an unexpanded nucleus treated with dilute MgC1, before FD from 20% ethanol. Shown are: nucleolus (Nu), chromatin (Ch), and contaminating cytoplasmic starch grains (3.Bar for both nuclei, 5 Fm. ~ 1 , 4 0 0 .
Fig. 9. High magnification stereo micrograph of a part of the expanded nucleus of the previous figure. The 18-65 nm nonfibrous and branched matrix material (MI, and the less common 12-18 nm chromatin fibers (F) make up most of the fine structure in this field of view. Also present are 35 nm granules ( G )in clusters, and large 200 nm nuclear bodies (Nb). Total tilt 16". Bar, 1.0 km. x 20,600.
the warmer temperature when structure might otherwise be destroyed. We often have successfully dried other kinds of grid-mounted specimens when the heater was set at -60°C; e.g., with fixed cells in culture, we observed very little damage or abnormal structure within cells near their thinner margins when they were frozen, warmed, and held for a time a s we
describe, at this higher temperature. This observation shows that these thinner parts were already dry before they reached the warmer temperature. The degree to which our apparatus satisfies all of the requirements thought to be necessary for proper drying is uncertain; however, a majority of our preparations are free of observable artifacts.
FREEZE-DRYING GRID-MOUNTED SPECIMENS
ACKNOWLEDGMENTS This work was supported in part by the U S . Department of Energy, Office of Health and Environmental Research, and by a grant to the senior author from the City University of New York PSC-CUNY Research Award Program, when he was a member of the Biology Department, Queens College, CUNY. Funds allocated to Jack Van’t Hof of the Biology Department, Brookhaven National Laboratory also contributed to the construction of the apparatus. Testing of prototypes began in 1980 when the senior author was on sabbatical in the Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder. The support of Mircea Fotino, the NIH high voltage EM users grant, and the assistance of George Wray was greatly appreciated. The authors also would like to thank Avril Woodhead for her suggestions during writing of the manuscript. REFERENCES Anderson, T. (1951) Techniques for the preservation of three-dimensional structures in preparing specimens for the electron microscope. Trans. N. Y. Acad. Sci. Ser. 11, 13:130-134. Chiovetti, R., McGuffee, L.J., Little, S.A., Wheeler-Clark, E., and Brass-Dale, J . (1987) Combined quick freezing, freeze-drying, and embedding tissue a t low temperatures and in low viscosity resins. J . Elec. Microsc. Tech., 5:l-15. Cole, A. (1967) Chromosome structure. In: Theoretical and Experimental Biophysics. Vol. 1. A. Cole, ed. Marcel Dekker, New York, pp. 305-375. Coulter, H.D., and Terracio, L. (1977) Preparation of biological tissues for electron microscopy by freeze-drying. Anat. Rec., 187:477-494. Coulter, H.D., and Terracio, L. (1978) An all glass freeze-dryer for TEM specimens with a n improved design for temperature regulation, fixation, and infiltration. 9th International Congress on Electron Microscopy, 11:60-61. DuPraw, E.J. (1970) DNA and Chromosomes. Holt, Rinehart and Winston, Inc., New York, 340 pp. Edelman, L. (1978) A simple freeze-drying technique for preparing
191
biological tissue without chemical fixation for electron microscopy. J . Microsc., 112:243-248. Elder, H.Y., Biddlecombe, W.H., Tetley, L., Wilson, S.M., and McEwan Jenkinson, D. (1986) Construction of low temperature freeze driers. EMSA Bull., 16:lll-113. Gall, J.G. (1966) Chromosome fibers by a spreading technique. Chromosoma, 20:221-223. Linner, J.G., Livesey, S.A., Harrison, D.S., and Steiner, A.L. (1986) A new technique for removal of amorphous phase tissue water without ice crystal damage: a preoperative method for ultrastructural analysis and immunoelectron microscopy. J. Histochem. Cytochem., 34:1123-1135. Mazia, D., Schatten, G., and Winfield, S. (1975) Adhesion of cells to surfaces coated with polylysine. J . Cell. Biol., 66:198-200. Menco, B., Ph. M. (1986) A survey of ultra-rapid cryofixation methods with particular emphasis on applications to freeze-fracturing, freeze-etching, and freeze-substitution. J. Elec. Microsc. Tech., 4: 177-240. Pawley, J . , and Ris, H. (1987) Structure of the cytoplasmic filament system in freeze-dried whole mounts viewed by HVEM. J. Microsc., 145:319-332. Phillips, T.E., and Boyne, A.F. (1984) Liquid nitrogen-based quick freezing: Experiences with bounce-free delivery of cholinergic nerve terminals to a metal surface. J . Elec. Microsc. Tech., 1:9-29. Ris, H. (1976) Levels of chromosome organization, Electron Micros. Proceedings of the 6th European Congress on Electron Microscopy, 1976, Vol. 11, pp. 21-25. Ris, H., Korenberg, J. (1979) Chromosome structure and levels of chromosome organization. In: Cell Biology A Comprehensive Treatise. Vol. 2. The Structure and Replication of Genetic Material. D.M. Prescott and L. Goldstein, eds. Academic Press, New York, pp. 267-361. Terracio, L., and Schwabe, K.G. (1981) Freezing and drying of biological tissues for electron microscopy. J. Histochem. Cytochem., 29: 1021-1028. Van’t Hof, J . (1975) DNA Fiber replication in chromosomes of a higher plant (Pzsurn satiuurn). Exp. Cell Res., 93:95-104. Woods, P.S., Ledbetter, M.C., and Van’t Hof, J . (1979) Use of EDTA in studies of isolated metaphase chromosome ultrastructure: visualization of a nonfibrous component. Proceedings of the 37th Annual EMSA Meeting, pp. 36-67. Woods, P.S., Van’t Hof, J., and Ledbetter, M.C. (1977) A differential centrifugation procedure for obtaining unstretched metaphase chromosomes from higher plants for transmission electron microscopy. Proceedings of the 35th Annual EMSA Meeting, pp. 402-403.
