A new method is offered for combined living cell and electron-microscopic studies of spermatocytes (or other cells) which normally do not adhere to glass. The key step is microinjection of glutaraldehyde near the target cell whenever desired during observation in life. Fixation begins and simultaneously the cell is stuck very firmly to the underlying coverslip. The method is easy and reliable: cells are almost never lost and are well preserved, except for membranes. The application of the method is illustrated by studies of micromanipulated grasshopper spermatocytes. A chromosome was detached from the spindle and placed in the cytoplasm. Before or after the beginning of chromosome movement back toward the spindle, the cell was fixed, sectioned and the manipulated chromosome observed in the electron microscope. If the detached chromosome had not moved by the time of fixation, no or only one or two microtubules were seen at its kinetochore, but if movement had occurred, a few microtubules were always present. The arrangement of these microtubules corresponded to the direction of movement, but they commonly were at an unusual angle relative to the kinetochore. The origin and role in chromosome movement of the microtubules seen near moving chromosomes far from the spindle is not yet established, but a speculation is offered. A goal for future work is the detailed analysis of the microtubules associated with individual moving chromosomes. Such an analysis is feasible because the moving chromosome is far removed from the confusing mass of spindle microtubules, and its value is enhanced because the direction of movement at the time of fixation is known.

Many questions about mitosis can be answered only through study of the same cell both in life and in the electron microscope. Methods already exist for correlated living cell/electron-microscopic studies of somatic mitosis in such excellent materials as Haemanthus endosperm (Molè-Bajer & Bajer, 1968) and Potorous tissue culture cells (e.g. Rattner & Berns, 1974; Roos, 1976). Similar studies of spermatocytes would be valuable. The large chromosomes and spindles of some spermatocytes make them as useful for observations on mitosis in living cells as the best somatic cells. In addition, spermatocytes offer features unique to meiotic cells such as a relatively great distance between opposed kinetochores (for the utility of this, see Nicklas, 1971, p. 268) and instructive deviations from orthodoxy (e.g. crane-fly sex chromosomes: Dietz, 1969; Forer & Koch, 1973). Unfortunately, 2 features of spermatocytes in culture make subsequent electron microscopy difficult. Spermatocytes do not adhere to the glass substrate and commonly are covered with a layer of inert oil —circumstances which almost always result in the loss of the desired cell during conventional preparation procedures for electron microscopy. The loss of cells can be overcome by preparing the cells in a fibrin clot (Forer, 1972), but our experience with the clot procedure has been marked by constant minor aggravations and inconsistent preservation of cell fine structure. A simple and reliable alternative procedure based on microinjection of fixative is described here, and the first results obtained by its use are presented.

Our immediate motivation to develop suitable electron-microscopic methods for spermatocytes was a desire to study micromanipulated chromosomes. Chromosomes can be detached from the spindle and placed wherever desired within the cell, and yet they invariably reassociate with the spindle and divide normally in anaphase (Nicklas, 1967). The experimental plan was to detach chromosomes from the spindle, to observe their behaviour up to the moment of fixation for electron microscopy, and finally to examine the micromanipulated chromosome in serial sections in the electron microscope. The preliminary but intriguing results obtained so far bear on the fine-structural meaning of ‘detachment’, on the origin of kinetochore microtubules (for definition, see legend, Fig. 2), and on the role of microtubules in chromosome motility.

Materials

Spermatocytes from 3 species were used for the fixation tests described in this section: the grasshoppers Dissosteira Carolina (L.) (from a North Carolina wild population) and Melanoplus differentialis (Thomas) (from a laboratory colony), and a cricket, Acheta domestica (L.) (from Flucker’s Cricket Farm, Baton Rouge, LA.). Spermatocytes from the grasshoppers were used for the studies reported in Results: Dissosteira for all the recent work, and Melanoplus for the older work by a different electron-microscopic method. The light- and electron-microscopic methods have also proved successful with spermatocytes from a mantispid neuropteran and several spiders.

Cell culture

The cell culture procedures were described earlier (Nicklas & Staehly, 1967, and references therein), but not in detail, and not with the saline now used.

The new saline is a PIPES-buffered formulation. The following concentrated stock solution is prepared fresh each month and stored in the refrigerator: 500 mM PIPES [piperazine-N, N’-bis(2-ethanesulphonic acid), Sigma], 27 mM KC1, 67 · 5 mM CaCl2, and 2 · 5 mM MgCl2 in glass-distilled water; the pH is adjusted to 6 · 8–6 · 9 with 10 N NaOH. On the day of use, 5 ml of the concentrated solution are diluted to between 27 and 38 ml (depending on the extent of evaporation during culture preparation and the physiological state of the animals; the required dilution usually does not change over a period of a week or two). Dilution to 32 ml is a good starting point. The proper dilution produces cells which are slightly hypotonic initially (as judged from the relatively vague appearance of metaphase chromosomes in phase-contrast microscopy), but which become isotonic in appearance after 2-4 h. This saline is a small but definite improvement over the one previously used (Nicklas & Staehly, 1967).

A convenient version of the cell culture chamber used on the inverted microscope (Nicklas & Staehly, 1976) for the present work is a stainless steel plate 1 × 29 × 76 mm in size with a 22-mm diameter hole in the centre, and a 25 × 25 mm coverslip sealed over the hole with petroleum jelly. Clean, wettable coverslips are essential and some brands are not satisfactory (we use ‘FisherBrand’, from Fisher Scientific, 3315 Winton Road, Raleigh, NC). Coverslips are cleaned by sonication in detergent, thoroughly rinsed (with sonication) in glass-distilled water, and stored in 95% ethanol. Shortly before use, coverslips are rinsed one at a time in glass-distilled water, inspected for wettability (retention of an even film of water during draining), then rinsed in 95% ethanol and dried by flaming. Dry coverslips remain wettable for only a few days. The coverslip is placed on the steel plate with 2 corners protruding to facilitate detaching it when desired. The chamber is used with the coverslip side down and the cells will lie on the upper surface of the coverslip, covered with halocarbon oil (‘Voltalef 10-S’: Ugine Kuhlmann, P.O. Box 159, Paramus, NJ, or 25 Boulevard de l’Amiral Bruix, Paris, 16è).

Cultures are prepared in a hood: any enclosed space containing a low-power dissecting microscope will serve. The relative humidity in the hood is raised to 70—90 % by boiling water just before a culture is prepared.

The testes are removed from an animal and quickly transferred into the diluted saline. The testes are viewed under a dissecting microscope and the fat is removed with watchmaker forceps. Four to 8 follicles are then picked up by the duct ends so that the opposite (‘distal’) ends hang downward. The group of follicles is then removed from the saline and drained of excess saline by dragging across a clean coverslip. Without delay, the distal ends are touched to the coverslip of a culture chamber, the follicles are cut with iridectomy scissors, and the duct ends are discarded. The distal ends are spread over the coverslip with watchmaker’s forceps, and the cells are quickly covered with Voltalef oil. Some practice is necessary to attain the speed and precision required in the critical period from the moment the follicles leave the saline until the cells are covered with oil. Successful preparations have numerous flat cells which remain healthy for a minimum of 6 h and occasionally for over 24 h. Healthy cells in division are unambiguously recognizable: the spindle is entirely free of visible inclusions other than chromosomes, is sharply defined by the mantle of mitochondria, and is longer pole-to-pole than across the equator. These are selected for further study (even in the best preparations, not all of the cells are healthy due to differential evaporation of the saline before the oil is added).

Light microscopy, micromanipulation, ciné recording, and analysis

All these procedures were carried out as previously described (Nicklas, 1963; Nicklas & Staehly, 1967), except that Agfa ‘Copex Pan Rapid ‘16-mm ciné film (sold in the U.S. as ‘VTE Pan’: H and W Co., St Johnsburn, VT.) was used.

Electron microscopy: fixation

The basis of our method is the discovery that microinjection of glutaraldehyde causes cells which formerly merely contacted a glass coverslip to adhere tightly to the glass. The secret is initial exposure of the cells to small volumes of glutaraldehyde solution so that fixation and attachment begin without displacing the cell from the coverslip. We use the term ‘microfixation’ to describe this process. After a few minutes of microfixation, the oil over the cells is removed and the coverslip is placed in a relatively large volume of fixative for ‘macrofixation’, followed by quite conventional OsO4 fixation, etc. Note that wherever exact dimensions, gauges, etc. are given in the following paragraphs, these are merely for the materials we happen to use. So long as the parts fit together the dimensions are not critical.

The general quality of fine structure preservation was judged from the density and richness of structural detail observed in the spindle, the cytoplasm, and in membranous organelles.

Apparatus

The equipment used is conventional and widely available, except for the micropipette holder. The injection apparatus consists of a 2-ml glass syringe with a spring-loaded piston, driven by a micrometer head (e.g. model 1208 of David Kopf Instruments, 7324 Elmo Street, Tujunga, CA.). The syringe is connected by tubing to the micropipette (hereafter, simply ‘pipette’). The pipettes are easily produced with the commercial ‘pipette pullers’ commonly found in neurophysiology laboratories (e.g. model 700C of David Kopf Instruments) or a rough but serviceable home-made device can be fabricated at nominal cost (details available upon request). Pipettes with an inside diameter at the tip of 0 · 5–2 μm are pulled from cleaned, soft-glass capillary tubing (0 · 9 mm outside, 0 · 7 mm inside diameter). The syringe and the pipette are connected with polyethylene tubing (PE-90 ‘Intermedic’ tubing 1 · 27 mm outside, 0 · 86 mm inside diameter, Clay Adams, 299 Webro Rd, Parsippany, NJ). A flange is produced at each end of the tubing by rolling the tubing between thumb and forefinger while holding the end near a lighted match. One flanged end is forced onto a 19-gauge hypodermic needle attached to the syringe, and the tubing and the barrel of the syringe are filled with silicone oil (50 centistoke ‘200 Silicone Fluid ‘, Dow Corning Chemicals, Midland, MI). The bulk of the air should be expelled, but a few bubbles in the syringe do not matter. The one novel element is the holder for the pipette, shown in Fig. 1, which is easier to use than the more elaborate conventional equivalents. After the pipette is connected to the tubing from the syringe (see Fig. 1, legend), the pipette holder is attached to a micromanipulator. Almost any available micromanipulator can be used because only coarse control of horizontal position is required, but control of vertical position with the sensitivity of a microscope fine-adjustment mechanism is desirable. We use a Brinkmann (Cantiague Rd, Westbury, NY) mechanical micromanipulator similar to their current model ‘CP’. The only requirements for the microscope are that it be inverted, preferably with a stage that is stationary during fine focusing, and that it be equipped with a long-working-distance condenser. The microscope and micromanipulator are placed on a steel plate resting on foam rubber for vibration absorption. The total cost of the micro-injection apparatus need not exceed S400.

Microfixation

A pipette of appropriate tip diameter (see below) is selected, and the shaft is scratched and broken to give a total length of about 20 mm. The pipette is loaded with fixative using a syringe equipped with a Swinny filter adapter (Millipore Corp., Bedford, MA; filter pore size: 0 · 22 μm) and a 26-gauge needle. The needle is inserted as far as possible into the shaft of the pipette, and fixative is forced into the shaft as the needle is slowly withdrawn. If the needle fits the shaft snugly, pipettes with a tip diameter over 1 μm are completely filled by this procedure; smaller pipettes retain air just behind the tip, which is expelled later. The pipette shaft is then inserted into the holder and connected to the tubing (see Fig. 1), with an air space 5–10 mm long separating the fixative from the silicone oil. Next, a cell culture reserved for preliminary adjustments is placed on the microscope. The pipette tip is lowered into the oil above the cells, brought into the field of view of a low-power objective, and the air at the tip, if any, is expelled. The tip is lowered to contact the aqueous layer around the cells, and the flow of fixative is regulated to a low rate, so that the flow stops whenever the tip is raised into the oil and begins anew the moment it is lowered to touch the aqueous layer. These characteristics of fixative flow, a happy consequence of the properties of water versus halocarbon oil, contribute notably to the ease of the whole procedure. The ‘adjustments culture’ is then removed, a previously selected cell in a second culture is brought into view, and the coverslip beneath it is scribed with a diamond marker concentric with the optical axis (e.g. Zeiss number 46 29 60).

At this point, the techniques diverge somewhat depending on the magnification at which the cells are viewed during fixation. Fixation while the pipette and the cell are viewed with a low-power objective (e.g. 16 times) is very easy and was used for the cells described in Results. For this variant, pipettes with an internal tip diameter of 1–2 μm are convenient. The tip of the pipette is placed in the oil above the cells, near the edge of the field of view. The target cell is observed with any desired objective during some interesting cellular process or micromanipulation, etc. When fixation is desired, the low-power objective is quickly swung into place, the pressure on the pipette is increased slightly, and the pipette is lowered to contact the aqueous layer 200 μm or so from the target cell. Only a small volume of fixative is allowed to flow before the pipette tip is raised back into the oil (the proper volume causes flow toward, but not to, the target cell; diffusion quickly does the rest). By moving the microscope stage, this process is repeated at 3 or more sites about equally distant from the target cell, thus surrounding the cell with fixative. The cell is now firmly stuck to the coverslip, and additional fixative is added closer to the cell until flow near, or even over, the cell is seen. The microfixation is now complete, 1 · 5–2 min after the first flow of fixative, but an additional 2 min or so elapse during the transfer to bulk fixative. The gap between the termination of observations at higher magnification and the beginning of fixation (cessation of chromosome movement) varies from 15 to 30s.

Alternatively, the cells may be fixed while they are viewed with the oil-immersion objective, thus eliminating any gap in the ciné record. The procedure used with the low-power objective is followed except that the rate of fixative flow is carefully adjusted to the minimum reliable rate just before observations are begun. Pipettes with a small tip diameter (0 · 5–1 /(m) are essential to achieve the very low rate of flow required. The pipette is placed in the oil above the cells at the very edge of the oil-immersion field of view. When fixation is desired, the pipette tip is lowered to the aqueous layer and then raised again almost immediately. The small volume of fixative released is so close to the target cell that chromosome movement ceases in 5 s or so. Then the low-power objective is rotated into position, the pressure in the injection syringe is increased, and larger volumes of fixative are applied 200 μm or so from the cell, exactly as in the first procedure.

The same pipette can be used to fix numerous cells over a period of several hours, if care is taken to prevent clogging the tip. The tip should be kept immersed in Voltalef oil (e.g. in the adjustments culture) when not in use, it should touch cells or debris (as opposed to the aqueous medium around them) only very briefly, and the fixative should always be under positive pressure when the tip touches cells or the aqueous medium.

Macrofixation

After microfixation, the Voltalef oil over the cells must be removed and the coverslip placed in bulk fixative. We have developed 2 procedures for removing the oil. In one, a large volume of fixative is injected beneath the oil using a 5-or 10-ml syringe fitted with a 19-gauge needle and filled with fixative. The cell chamber is placed on a flat surface to support the coverslip, the bevelled edge of the syringe needle is placed flat against the coverslip on the side with the cells, 10 mm or so from the target cell, and fixative is gently expelled until it forms a dome rising above the chamber. The bulk of the oil around the edge of the dome is removed with cotton applicator sticks, the covershp is detached from the chamber, the side with the cells is flushed gently but thoroughly with fixative, and finally the coverslip is placed in a small coplin jar full of fixative. The coverslip is handled at angles which favour the rapid removal of most of the oil with minimal contamination of the area near the target cell.

Alternatively, the Voltalef oil can be removed after microfixation by placing the culture chamber in Freon 22 (Rebhun, 1965) cooled to its freezing point with liquid nitrogen. The oil solidifies, the coverslip is detached from the chamber while in the Freon 22, and then the coverslip is transferred to bulk fixative at room temperature. The oil over the cells is completely removed by the freezing procedure. General cell structure is at least as well preserved as in the injection procedure, and most significantly, membranes are better preserved (possibly because no film of oil is left behind). Unfortunately, the freezing procedure in its present form is often associated with fragmentation and loss of microtubules and cannot be recommended at this time.

Fixative solutions and timing

Throughout this study purified glutaraldehyde (Ladd, P.O. Box 901, Burlington, VT.) from sealed glass ampoules was diluted and used the same day.

The micromanipulated cells in the recent experiments described below (e.g. Figs. 3-5, pp. 96 and 98) were fixed according to the following schedule using ‘phosphate buffer’ (10mM Sorensen’s phosphate buffer, 0 · 5 % NaCl, pH 6 · 7):

  1. Microfixation: 6% glutaraldehyde in phosphate buffer, 4 · 5–5 · 5 min.

  2. Macrofixation: 3 % glutaraldehyde in phosphate buffer, 30 min.

  3. Rinse: phosphate buffer, 4 changes, 30 min total time.

  4. Second fixation: 2 % OsO4 in phosphate buffer: 60 min.

  5. Rinse: distilled water, 2 changes, 20 min total time.

The minimal loss of spindle birefringence was the criterion used to select the concentration of glutaraldehyde for microfixation. Spindle birefringence (strictly, the retardation) was measured in the same cell first in life and then again in the rinse following glutaraldehyde fixation (OsO4 treatment produces so much birefringence in membranous organelles that later measurements are meaningless). The birefringence was measured with a Nikon rectified polarization microscope (Inoué & Hyde, 1957) using a Brace-Köhler compensator (e.g. Bennett, 1950, pp. 652654, ‘full compensation’ method). Each value obtained is an average between chromosomal fibre and background birefringence for regions near the thickest part of the spindle. Measurements were made on two or more cells after microfixation in 3, 6, or 9 % glutaraldehyde. The drop in birefringence is 20–30 % after 6 % glutaraldehyde, but it is 50–60 % for the other 2 concentrations. Six per cent is higher than the conventional glutaraldehyde concentration, but the necessity for this is no surprise because the concentration at the target cell is lowered by dilution and diffusion during microinjection.

Recently, the preservation of spindle birefringence has also been used to screen alternative fixative and rinse solutions and to select the optimal duration of the initial steps in the schedule. The best procedure involves the use of PIPES buffer and shorter durations (Snyder & McIntosh, 1975, and personal communication). This procedure is not a striking improvement over the phosphate buffer schedule, but may be marginally better, since the birefringence loss is less for some cells (range: 10–25%) and the preservation of fine structure is at least as good. The schedule is modified as follows: for ‘phosphate buffer’, substitute 10mM PIPES, 0 3% NaCl, pH 6 · 8–6 · 9, and change the durations to 3 ± 1 min, 15 min, 3 min, 30 min, and 10 min for steps 1 to 5 respectively. An example is shown in Fig. 2, p. 94. The values given for the drop in birefringence after phosphate-or PIPES-buffered glutaraldehyde apply to Dissos-teria, Melanoplus, and Acheta spermatocytes. Numerous additional fixatives have been tried on Melanoplus and/or Acheta spermatocytes and have been rejected because the loss of birefringence is often or always greater than 30 %. The rejected fixatives are glutaraldehyde in media based on a tubulin-assembly medium (Luftig, McMillan, Weatherbee & Weihing, 1977) or a microtubule-stabilizing medium (Filner & Behnke, 1973); glutaraldehyde in phosphate or PIPES buffer plus dimethylsulphoxide, or plus TAPO (Reedy, 1976), or with sucrose instead of NaCl; and a combined glutaraldehyde/OSO4 fixative (Fristrom & Fristrom, 1975).

Electron microscopy: later steps

After the rinse following OsO4, the coverslips were processed conventionally through a graded ethanol series to propylerie oxide, infiltrated in propylene oxide-resin mixtures (Epon-Araldite: Mollenhauer, 1964), and embedded in Epon-Araldite in a Chang mould (Chang, 1971). In transferring coverslips, speed is essential to prevent drying. When necessary, processing was suspended overnight during infiltration, but not at any earlier stage. The resin was polymerized at 37 °C until tacky and then at 6oºC. After partial polymerization, each plastic block was separated from the mould, the diamond-scribed circle on the coverslip located, and a corresponding mark made on the back side of the block. The coverslip was removed either by freezing partly polymerized blocks (Branson, 1971) in liquid nitrogen (example: Figs. 4 and 5, pp. 96 and 98), or by floating the block, coverslip side down, on concentrated hydrofluoric acid (Moore, 1975) (example: Fig. 2, p. 94) after 2 days or so at 60°C. The latter method is highly recommended. As Moore says, the hydrofluoric acid treatment causes absolutely no detectable changes in cell fine structure or in the sectioning properties of the plastic: no cell need ever again be lost at this stage. The blocks were then returned to the oven to complete polymerization for a total of 4–6 days at 60°C. The blocks were trimmed, and serial sections were cut, mounted, and stained as previously described (Brinkley, Murphy & Richardson, 1967; Kubai, 1973). The sections were examined either with a Siemens 101 or a Philips 201 electron microscope operating at 80 or 60 kV respectively.

Bivalents on the spindle

Bivalents in the normal orientation on a metaphase spindle are shown in Fig. 2 as a reminder of the arrangement of kinetochore microtubules at the first meiotic division, and also as an illustration of the results attained with the PIPES-buffer, short-fixation schedule. Some necessary definitions are given in the legend. Note that each group of kinetochore microtubules extends straight away from a kinetochore toward one pole: most kinetochore microtubules are nearly parallel to the interkinetochoric axis. In Dissosteira, as in all orthopterans examined to date, the kinetochore is a more (Fig. 5 A), or less (Fig. 2), distinct ball set in a shallow depression, and thus is similar to the ‘ball and cup’ kinetochore of Haemanthus (Bajer & Molè-Bajer, 1972). The kinetochore(s) of each half-bivalent might be single or double (two closely appressed, sister kineto-chores). The latter is far more likely (e.g. Müller, 1972), but because no hint of doubleness is visible here, the singular will be used solely to avoid apparent contradiction between text and figures. The kinetochore is at the apparent end of every halfbivalent at metaphase I in Dissosteira — the very short second arm is not visible. The exact number of kinetochore microtubules is not known, but 30 per metaphase halfbivalent is a reasonable estimate for materials fixed by the current methods.

Manipulated bivalents

A healthy, relatively flat cell in prometaphase or metaphase I was selected, and using a fine glass needle, a bivalent was detached and placed in the cytoplasm as far from the spindle as possible. In every experiment, a mechanical criterion for detachment was satisfied: the bivalent could be freely moved anywhere in the cell without the least sign of the attachment to the spindle which is so conspicuous before detachment (Nicklas & Staehly, 1967; Nicklas, 1967). The cells were fixed either before or during the movement of the manipulated chromosome back toward the spindle.

Recent studies

An example is shown in Figs. 3-5. In this instance, a bivalent was detached and then allowed to move almost all the way back to the spindle before fixation. The 0 ·0-min print (lower left) of Fig. 3 shows the bivalent after reorientation following a previous detachment. The bivalent was then redetached and placed as shown in the 1 ·1-min print (the needle was removed at 0 · 9 min). Movement was reinitiated 5 s after the needle was removed, and the bivalent moved straight back toward the spindle to a position much nearer to the equator than to either pole (1 ·1-to 2 ·1-min prints). The inner half-bivalent led the way in this movement, but the outer half-bivalent showed an independent rotation toward the spindle (1 · 1- and 1 · 6-min prints). Thus the movement was consistent with forces acting at the kinetochore of each half-bivalent. Electron-microscopic observations disclose microtubules at both kinetochores. The inner kinetochore lies amongst numerous microtubules generally parallel with those of the spindle. The microtubules most intimately associ-ated with the inner kinetochore are shown in Fig. 4: several extend to the left (A-C) and right (D, E). Close inspection will disclose that at least some of these microtubules enter the kinetochore (see especially Fig. 4c), and while some may terminate there, others pass on through. The disposition of these microtubules is variable, but as a group, those on the left run toward the left-hand pole, while those on the right diverge toward the cytoplasm at an angle of c. 20° to the spindle axis. Thus neither group of microtubules runs in the direction of this half-bivalent’s movement as observed before fixation. Most significant is the angle of these microtubules relative to the kinetochore: all are roughly perpendicular to the inter-kinetochoric axis, and none is parallel with this axis, unlike the normal metaphase-anaphase orientation (Fig. 2). The microtubules associated with the outer kinetochore (Fig. 5) are especially interesting. They run from the periphery of the kinetochore (A, B), or beyond (B, arrowhead), toward the spindle, and generally in the direction of movement of the outer half-bivalent. Some of these microtubules can be traced for several micrometres (e.g. Fig. 3B, the microtubule identified by arrows) but unfortunately their association with spindle microtubules has not been documented. Here again, all the observed microtubules are perpendicular, or nearly so, to the end of the chromosome, a pattern utterly different from that of the kinetochore microtubules seen in Fig. 2. Most or all of the microtubules visible in Fig. 5 are very likely associated with the manipulated chromosome (1) because their angles are not random and are related to the direction of movement of the halfbivalent, and (2) because, in the absence of a moving chromosome, no similarly high concentration of microtubules has ever been seen in comparable cytoplasmic areas of several cells examined (for example, for the cell shown in Figs. 3-5, a total of 11 μm of microtubule length was found in 320 μm2 of cytoplasmic area near, but not at, the moving chromosome —an average of 0 · 3 μm per 10 μm2 —while in the area shown in Fig. 5 A, the value is about 2 · 9 μm per 10 μm2).

Three additional cells from the same experimental series have been examined. The first of these confirms in every respect the results shown in Figs. 3-5. The bivalent had moved close to the spindle before fixation, and both half-bivalents showed independent movement similar to that in Fig. 3. The disposition of microtubules at the inner and outer kinetochores was nearly identical to that seen in Figs. 4 and 5. In the other 2 cells, the detached bivalent was allowed to move only half way back to the spindle before fixation, and only one half-bivalent showed independent movement. The kinetochoric end of one half-bivalent (the ‘inner’) led the way, the movement was toward the equatorial region of the spindle, and the other (‘outer’) half-bivalent showed only passive movement. Electron-microscopic observations were fragmentary due to loss of sections. From the vicinity of the inner kinetochore in both examples, one or two microtubules ran toward each spindle pole; those running toward one pole certainly entered the kinetochore but this could not be documented for those running toward the other pole. At the other (outer) kinetochores, one or two short segments of microtubules were seen in one example (4 sections), and none in the other (complete serial series). The data are scanty, but these kinetochores of detached half-bivalents that had not actively moved when fixed, clearly did not display as many microtubules as those that had moved (Figs. 4, 5), let alone the number found before detachment (Fig. 2).

Older studies

A now superseded procedure involving freezing the cells before fixation was used for electron microscopy of a manipulated bivalent or X-chromosome in 10 Melanoplus differentialis spermatocytes (Brinkley & Nicklas, 1968, and unpublished). Preservation was highly variable in quality and never as good as that consistently obtained now. However, the consistency of the findings with the more recent results should be noted. Considered in terms of individual half-bivalents (or the unpaired X-chromosome), 11 had moved by the time of fixation, and without exception a few microtubules were found at each kinetochore. All 11 were in bivalents which had been placed relatively near the spindle after detachment, and all were moving more nearly toward one pole rather than directly toward the equator. In every case the kinetochoric microtubules lay in the direction of chromosome movement. In addition, 5 half-bivalents which had not shown independent movement when fixed were observed. For 4 of these, no microtubules were seen at the kinetochore and the fifth had only a single microtubule. The most informative single example has been illustrated previously (Nicklas, 1971, fig. 18). In this instance, when fixation occurred, one half-bivalent had just begun to swing toward the spindle, while the other had not moved. Three microtubules were found at the kinetochore of the moving halfbivalent, and these ran perpendicular to the interkinetochoric axis as in the examples already considered. No microtubules were found at the kinetochore of the nonmoving bivalent. A complete set of serial sections through the kinetochores of both half-bivalents was examined.

Methods

We recommend the glutaraldehyde microinjection method as the one of choice for combined living-cell/electron-microscopic studies, not only for spermatocytes but for all cells which do not normally adhere tightly to glass. The apparatus and techniques required may seem formidable, but in fact the apparatus unique to the method is inexpensive and the techniques can be mastered in a week. The method has been in use for over a year and a variety of materials have been observed in the electron microscope after diverse experimental and descriptive studies in life. Fewer than 5% of the cells were lost during processing and none during removal of the coverslip, when Moore’s (1975) hydrofluoric acid procedure was used. The only significant sources of grief were occasional losses during trimming the plastic block and in serial sectioning —problems common to all methods. Qualitatively good preservation of microtubules, chromosomes, and general cell architecture was consistently obtained. However, membranes are commonly less sharply defined than after standard procedures. The problem of membrane preservation is under study, and an early solution seems possible since the source of the problem has been identified (see the remarks on removal of the oil over the cells by freezing after microfixation, page 91).

At best, only 70–90% of the spindle birefringence remains after glutaraldehyde fixation, and this may well be a general phenomenon, not limited to our materials or methods. Sato, Ellis & Inoue (1975) have shown that the birefringence of isolated spindles can be preserved by adding hexylene glycol to the fixative and other solutions, but no equally reliable method is available for whole cells. Two previous studies on the preservation of spindle birefringence in whole cells are on record. LaFountain (1974, 1976) compared spindle birefringence in living crane-fly spermatocytes with that remaining after glutaraldehyde fixation. He found a 50% drop in birefringence for cells in metaphase, but no drop for cells in anaphase. McIntosh, Cande, Snyder & Vanderslice (1975) measured spindle birefringence in cultured mammalian cells during the first 10–12 min of glutaraldehyde fixation and found substantial loss in some cells but none in others. Most importantly, they present data showing that this initial reduction in birefringence is not associated with a decrease in the number of spindle microtubules. Our concern remains, however, because the total loss of birefringence is unknown, since measurements after OsO4 are not meaningful, as noted above. If any part of the total loss were due to microtubule disarray or depolymerization, a particular class of microtubules might well be differentially affected or even missing altogether (for literature and observations on differential microtubule lability, see e.g. Brinkley & Cartwright, 1975; Rieder & Bajer, 1977). Certainly the gross variation in kinetochore microtubule number reported for the spermatocytes of a single species of crane fly (LaFountain, 1976; Forer & Brinkley, 1977) is disquieting, even though the ratios between classes of microtubules may be unaffected (Forer & Brinkley, 1977). Measurement of spindle birefringence before and after glutaraldehyde fixation is an incomplete and ambiguous guide to quantitatively good preservation of spindle fine structure, but it is the only available indicator. We urge such measurements as a prelude to any quantitative or semiquantitative study of spindle microtubules, especially if the inferences drawn depend on the preservation of representative numbers and lengths of all classes of microtubules.

Chromosome reattachment to the spindle

Some significant fine-structural features of reattachment are now evident, despite the rather fragmentary evidence. The necessary background information gained through studies of living cells is as follows. (1) ‘Detachment’ of chromosomes from the spindle involves the complete loss of the prior chromosome-to-spindle connexions: reattachment is not biased in favour of reassociation with the pole to which a given half-bivalent was originally attached (Nicklas, 1967) nor even with the same spindle (Nicklas, 1977). (2) Reattachment never fails to occur: no matter how far from the spindle a detached bivalent is placed, it very soon begins to move back toward the spindle, one kinetochoric end foremost, and at a relatively high speed (Nicklas, 1967). (3) For detached bivalents placed near or within the spindle, the subsequent movement may be toward a pole (Nicklas, 1967), but contrary to Nicklas (1967), Begg (1975) is entirely correct that bivalents placed far from the spindle almost invariably move first toward the equatorial region of the spindle.

The electron-microscopic observations to date can be summarized as follows, in terms of individual half-bivalents, the unit of independent motility. (1) If a given halfbivalent had not begun independent movement by the time of fixation, usually no, at most two, microtubules were seen at its kinetochore. (2) If, however, movement had begun, microtubules were always seen at the kinetochore. (3) In the simpler examples, these microtubules extended toward one part of the spindle (Fig. 5), always in the same direction as the chromosome movement observed before fixation. (4) Whether the relationship between movement and microtubule disposition was simple or complex (discussed below), the microtubules at the kinetochore were often perpendicular to the interkinetochoric axis, and sometimes ran toward opposite poles. This is in striking contrast to the orientation of microtubules with respect to normal halfbivalents in metaphase or anaphase. A major gap in our observations should be emphasized: the distal portions of kinetochore microtubules have not yet been observed for bivalents which were still some distance from the spindle when fixed. It is not yet known where these microtubules may abut spindle microtubules, and at what angles.

Microtubules were invariably seen at the kinetochores of moving half-bivalents and yet were absent or reduced in number at the kinetochores of non-moving half-bivalents in the same set of experiments. In fact, even the opposed kinetochores in the same bivalent can differ in this fashion (e.g. Nicklas, 1971, fig. 18). Therefore, in this case, the absence of microtubules surely is meaningful: either kinetochore microtubules are actually not present after detachment and before renewed movement, or at least they are less easily preserved than normally. More observations on well preserved cells are necessary here. The converse observation, of an association of microtubules with renewed chromosome movement, is hardly a surprise, but the results provide some experimental evidence for the assumption that the direction of movement is determined by the disposition of microtubules. Thus, forces parallel to the microtubule axes would produce the movement observed in the simpler examples of movement after detachment, as in normal prometaphase and anaphase. This adds to the earlier evidence (Nicklas, 1967) suggesting that the movement after detachment from the spindle is produced by orthodox mitotic mechanisms in an unorthodox situation.

We now consider the more complex examples. These involve bivalents placed far from the spindle in which the half-bivalent closer to the spindle moved toward the equator. If the cell was fixed as the bivalent neared the spindle, an arrangement of microtubules running toward both poles was seen at the kinetochore of the halfbivalent closer to the spindle. The same arrangement may exist at earlier stages of movement back toward the spindle, but the evidence is incomplete. Preliminary evidence for a similar bipolar configuration of microtubules in unmanipulated meiotic cells has been reported (Luykx, 1965; Wagenaar & Bray, 1973), and in multipolar somatic mitoses, an occasional single sister kinetochore with microtubules running toward two different poles has been documented beyond doubt (Heneen, 1975; Lambert & Bajer, 1975). All these configurations, including those reported here, normally are probably short-lived intermediates in processes leading to the orthodox metaphase orientation, as östergren (1951) suggested long ago for similar, if simpler, intermediates. We have no evidence to sustain any conjectures about the final transition from, e.g. the microtubule arrangement in Fig. 4 to that in Fig. 2, but we cannot resist a brief speculation on the origin and immediate consequences of bipolar microtubule arrangements in detached chromosomes. Consider the possible events at a halfbivalent such as the one facing the spindle in the 1 ·1-min print, Fig. 3. Suppose (1) that immediately after detachment, the half-bivalent has no or only very short fragments of kinetochore microtubules, (2) that movement is reinitiated only when microtubules are present at the kinetochore, usually oriented toward both spindle poles, and (3) that microtubules running all the way back to the spindle are essential for movement (either single or interconnected microtubules span the whole distance). We suggest that such microtubules originate from pre-existent spindle —probably astral —microtubules which either bind directly to the kinetochore or form lateral linkages with short (pre-existent or freshly nucleated) kinetochore microtubules. This would explain (1) the very rapid appearance of long microtubules at the kinetochore following detachment, (2) the angle of these microtubules (usually nearly perpendicular to the kinetochore but radial relative to the asters), and (3) the direction of movement seen by Begg (1975). The rationale behind the third point is that microtubules toward both poles, if associated with orthodox mitotic force production parallel to the microtubule axes, would lead to forces toward both poles on the moving half-bivalent. The net force would then be more nearly toward the spindle equator than toward either one of the 2 poles (this explanation was first suggested to us by Mr Steven Cohn, Washington University, St Louis).

Speculation aside, 2 conclusions can be drawn directly from the observations. First, it is evident that the initial events in kinetochore-spindle interaction after detachment are more complex —and hence more interesting —than one of us has imagined (Nicklas, 1967, 1974). Second, detached chromosomes furnish unique material for electron-microscopic studies both on the origin of kinetochore microtubules and on the microtubule-microtubule associations involved in the movement of a single chromosome. The uniqueness derives from the possibility of placing a detached chromosome at any desired angle to, and distance from, the spindle. Thus the subsequent movement can be made to occur in an area of the cell so impoverished in microtubules relative to the spindle that the few that are present may be completely analysable and unambiguously associated with the particular movements executed by one chromosome.

We are grateful to Dr Dwayne Wise for contributing ideas and practical experience to the development of the techniques and for the drawing, Fig. 1. Dr Melvin Lieberman generously gave advice and access to equipment for pipette fabrication and Dr Michael K. Reedy provided electron-microscope facilities. For skilled technical assistance, Joiner Cartwright, Jr, Tom Hays, and Carol Koch deserve special thanks. These studies were supported in part by research grants GM-13745 from the Division of General Medical Sciences, DHEW; CA-22610 from the National Cancer Institute, DHEW; and PCM 76-20131 from the National Science Foundation.

Bajer
,
A. S.
&
Molè-Bajer
,
J.
(
1972
).
Spindle dynamics and chromosome movements
.
Int. Rev. Cytol
., uppl.
3 (ed
.
G. H.
Bourne
&
J. F.
Danielli
).
New York and London
:
Academic Press
.
Begg
,
D. A.
(
1975
).
Chromosome movement following detachment from the meiotic spindle by micromanipulation
.
J. Ceil Biol
.
67
,
25
a (abstr.).
Bennett
,
H. S.
(
1950
).
The microscopical investigation of biological materials with polarized light
.
In McClung’s Handbook of Microscopical Technique
(ed. R.
McClung
Jones
), pp.
591677
.
New York
:
Hoeber
.
Branson
,
S. H.
(
1971
).
Epon-embedded cell monolayers. Separating partially polymerized blocks from glass surfaces
.
Expl Cell Res
.
65
,
253
256
.
Brinkley
,
B. R.
&
Cartwright
,
J.
Jr
(
1975
).
Cold-labile and cold-stable microtubules in the mitotic spindle of mammalian cells
.
Ann. N. Y. Acad. Sci
.
253
,
428
439
.
Brinkley
,
B. R.
,
Murphy
,
P.
&
Richardson
,
C.
(
1967
).
Procedure for embedding in situ selected cells cultured in vitro
.
J. Cell Biol
.
35
,
279
283
.
Brinkley
,
B. R.
&
Nicklas
,
R. B.
(
1968
).
Ultrastructure of the meiotic spindle of grasshopper spermatocytes after chromosome micromanipulation
.
F. Cell Biol
.
39
,
16a
17a
(abstr.).
Chang
,
J. P.
(
1971
).
A new technique for separation of coverglass substrate from epoxyembedded specimens for electron microscopy
.
J. Ultrastruct. Res
.
37
,
370
377
.
Dietz
,
R.
(
1969
).
Bau und Funktion des Spindelapparats
.
Natuncissenschaften
56
,
237
248
.
Filner
,
P.
&
Behnke
,
O.
(
1973
).
Stabilization and isolation of brain microtubules with glycerol and dimethlysulfoxide (DMSO)
.
J. Cell Biol
.
59
,
997
(abstr.).
Forer
,
A.
(
1972
).
A method for making preparations of living crane fly spermatocytes for study with light microscopy followed by electron microscopy
.
Cytobiologie
6
,
403
409
.
Forer
,
A.
&
Brinkley
,
B. R.
(
1977
).
Microtubule distribution in the anaphase spindle of primary spermatocytes ofacrane fly (Nephrotoma suturalis)
.
Can. J. Genet. Cytol
.
19
,
503
519
.
Forer
,
A.
&
Koch
,
C.
(
1973
).
Influence of autosome movements and of sex-chromosome movements on sex-chromosome segregation in crane fly spermatocytes
.
Chromosoma
40
,
417
442
.
Fristrom
,
D.
&
Fristrom
,
J. W.
(
1975
).
The mechanism of evagination of imaginal disks of Drosophila melanogaster. I. General considerations
.
Devi Biol
.
43
,
1
23
.
Heneen
,
W. K.
(
1975
).
Kinetochores and microtubules in multipolar mitosis and chromosome orientation
.
Expl Cell Res
.
91
,
57
62
.
Inoué
,
S.
&
Hyde
,
W. L.
(
1957
).
Studies on depolarization of light at microscope lens surfaces. II. The simultaneous realization of high resolution and high sensitivity with the polarizing microscope
.
J. BLophys. BLochem. Cytol
.
3
,
831
838
.
Kubai
,
D. F.
(
1973
).
Unorthodox mitosis in Trichonympha agilis: kinetochore differentiation and chromosome movement
.
F. Cell Sci
.
13
,
511
552
.
Lafountain
,
J. R.
, Jr
(
1974
).
Birefringence and fine structure of spindles in spermatocytes of Nephrotoma suturalis at metaphase of first meiotic division
.
J. Ultrastruct. Res
.
46
,
268
278
.
Lafountain
,
J.R.
, Jr
(
1976
).
Analysis of birefringence and ultrastructure of spindles in primary spermatocytes of Nephrotoma suturalis during anaphase
.
J. Ultrastruct. Res
.
54
,
333
346
.
Lambert
,
A.-M.
&
Bajer
,
A.
(
1975
).
Fine structure dynamics of the prometaphase spindle
.
J. Microscopie Biol. cell
.
23
,
181
194
.
Luftig
,
R. B.
,
Mcmillan
,
P. N.
,
Weatherbee
,
J. A.
&
Weihing
,
R. R.
(
1977
).
Increased visualization of microtubules by an improved fixation procedure
.
J. Histochem. Cytochem
.
25
,
175
187
.
Luykx
,
P.
(
1965
).
Kinetochore-to-pole connections during prometaphase of the meiotic division in Urechis eggs
.
Expl Cell Res
.
39
,
658
668
.
Mcintosh
,
J. R.
,
Cande
,
Z.
,
Snyder
,
J.
&
Vanderslice
,
K.
(
1975
).
Studies on the mechanism of mitosis
.
Am. N. Y. Acad. Sci
.
253
,
407
427
.
Molè-Bajer
,
J.
&
Bajer
,
A.
(
1968
).
Studies of selected endosperm cells with the light and electron microscope. The technique
.
Cellule
67
,
257
265
.
Mollenhauer
,
H. H.
(
1964
).
Plastic embedding mixtures for use in electron microscopy
.
Stain Technol
.
39
,
111
114
.
Moore
,
M. J.
(
1975
).
Removal of glass coverslips from cultures flat embedded in epoxy resins using hydrofluoric acid
.
J. Microscopy
104
,
205
207
.
Müller
,
W.
(
1972
).
Elektronenmikroskopische Untersuchungen zum Formwechsel der Kinetochoren während der Spermatocytenteilungen von Pales ferruginea (Nematocera)
.
Chromosoma
38
,
139
172
.
Nicklas
,
R. B.
(
1963
).
A quantitative study of chromosomal elasticity and its influence on chromosome movement
.
Chromosoma
14
,
276
295
.
Nicklas
,
R. B.
(
1967
).
Chromosome micromanipulation. II. Induced reorientation and the experimental control of segregation in meiosis
.
Chromosoma
21
,
17
50
.
Nicklas
,
R. B.
(
1971
).
Mitosis
.
In Advances in Cell Biology
, vol.
2
(ed.
D. M.
Prescott
,
L.
Goldstein
&
E.
Mcconkey
), pp.
225
297
.
New York
:
Appleton-Century-Crofts
.
Nicklas
,
R. B.
(
1974
).
Chromosome segregation mechanisms
.
Genetics, Princeton
78
,
205
213
.
Nicklas
,
R. B.
(
1975
).
Chromosome movement: current models and experiments on living cells
.
In Molecules and Cell Movement
(ed.
S.
Inoué
&
R. E.
Stephens
), pp.
97
117
.
New York
:
Raven Press
.
Nicklas
,
R. B.
(
1977
).
Chromosome distribution
:
experiments on cell hybrids and in vitro
.
Phil. Trans. R. Soc. Ser. B
277
,
267
276
.
Nicklas
,
R. B.
&
Staehly
,
C. A.
(
1967
).
Chromosome micromanipulation. I. The mechanics of chromosome attachment to the spindle
.
Chromosoma
21
,
1
16
.
Ostergren
,
G.
(
1951
).
The mechanism of co-orientation in bivalents and multivalents. The theory of orientation by pulling
.
Hereditas
37
,
85
156
.
Rattner
,
J. B.
&
Berns
,
M. W.
(
1974
).
Light and electron microscopy of laser microirradiated chromosomes.
.
Cell Biol
.
62
,
526
533
.
Rebhun
,
L. I.
(
1965
).
Freeze-substitution: fine structure as a function of water concentration in cells
.
Fedn Proc. Fedn Am. Socs exp. Biol
.
24
,
suppl
. 15,
S217
S232
.
Reedy
,
M. K.
(
1976
).
Preservation of X-ray patterns from frog Sartorius muscle prepared for electron microscopy
.
Biophys. J
.
16
,
126
a (abstr.).
Rieder
,
C.
&
Bajer
,
A. S.
(
1977
).
Heat-induced reversible hexagonal packing of spindle microtubules
.
F. Cell Biol
.
74
,
717
725
.
Roos
,
U.-P.
(
1976
).
Light and electron microscopy of rat kangaroo cells in mitosis. III. Patterns of chromosome behaviour during prometaphase
.
Chromosoma
54
,
363
385
.
Sato
,
H.
,
Ellis
,
G. W.
&
Inoué
,
S.
(
1975
).
Microtubular origin of mitotic spindle form birefringence. Demonstration of the applicability of Wiener’s equation
.
J. Cell Biol
.
67
,
501
517
.
Snyder
,
J. A.
&
Mcintosh
,
J. R.
(
1975
).
Initiation and growth of microtubules from mitotic centers in lysed mammalian cells
.
J. Cell Biol
.
67
,
744
760
.
Wagenaar
,
E. B.
&
Bray
,
D. F.
(
1973
).
The ultrastructure of kinetochores of unpaired chromosomes in a wheat hybrid
.
Can. J. Genet. Cytol
.
15
,
801
806
.