counting and measuring imps and pits: why accurate counts are exceedingly rare · 2008-11-04 ·...

JOURNAL OF ELECTRON MICROSCOPY TECHNIQUE 13204-215 (1989)

Counting and Measuring IMPs and Pits: Why Accurate Counts Are Exceedingly Rare JOHN E. RASH AND F. DENNIS GIDDINGS Department of Anatomy and Neurobiology (J.E.R., F.D.G.), Cell and Molecular Biology Program (J.E.R.) and Program in Neuronal Growth and Development (J.E.R., F.D.G.), Colorado State University, Ft. Collins, Colorado 80523

KEY WORDS Complementarity, Freeze-fracture replica, Pre-fracture, Stress-fracture, Water vapor contamination

ABSTRACT Particle counting and measuring techniques are now widely used to characterize normal membranes and to identify molecular changes occurring during development, maturation, and aging during progression of disease and following pharmacological manipulation. However, the use of particle counting and measuring for the identification of molecular changes in membranes has been premature. We show that current procedures rarely yield replicas that are free of cryogenic or mechanical prefractures, and as a result, the “complementarity” of membrane faces is severely compromised. However, with simple alterations of procedure, combined with the resolve to recognize and discard images of pre-fractured membrane faces, a high degree of “complementarity” may be obtained. Criteria for recognizing the occurrence and relative frequency of noncomplementarity are presented and a cleaving method for avoiding a primary source of water vapor contamination is described. In such replicas, membrane pits are found in equivalent numbers and near-identical diameters as the intramembrane particles (IMPs) in the complementary-type membrane faces. When conditions of “cold fracture” and immediate replication are demonstrated, fracture faces are minimally contaminated by frozen water vapor, yielding images where 1) diameters of IMPs vs. pits are very nearly identical, 2) large diameter IMPs are very rare, and 3) the numbers of IMPs and pits are increased substantially over the numbers currently reported. Thus, we reiterate previous proposals that complementarity of membrane faces is the single most important criterion that must be met before accepting the validity of IMP counts or for attributing perceived changes in IMP density or size to conditions of experimental manipulation, to normal developmental processes, or to disease etiology.

INTRODUCTION If IMPS (intramembrane particles) reflect the mor-

phological correlate for membrane integral proteins (Singer and Nicolson, 1972), it would be reasonable to expect that the numbers, sizes, and morphologies of IMPs in replicas of similar membranes should be very nearly identical from experiment to experiment and from laboratory to laboratory. Moreover, it would be reasonable to expect specific classes of membrane integral proteins to exhibit characteristic sizes, shapes, and cleaving patterns reflecting their molecular weights and/or characteristic states of aggregation. Based on these two assumptions, there are numerous published studies suggesting that there are characteristic numbers of IMPs per unit area in each type of membrane and that these values are altered in statistically sig- nificant amounts as a direct reflection of membrane alterations associated with development, maturation, and aging; during progression of various disease states; and following pharmacological manipulation. How- ever, direct comparison of the published data reveals an extreme lack of reproducibility in IMP counts and measurements, even when the counts were obtained from ostensibly similar or identical membranes. For example, in freeze-fracture studies comparing the number of dispersed IMPs in normal patients with

those in patients with Duchenne muscular dystrophy, the number of IMPs is reported either 1) to decrease significantly (Bonilla et al., 1981; Schotland, 1977; Schotland et al., 1977) or 2) to increase significantly (Ketelsen, 1975; Yoshioka and Okuda, 1977) or 3) to be unaltered (Rash et al., 1979b; see also Osame et al., 1981, for IMP counts from dystrophic muscle cells in culture). More distressing are the actual data, with reported values for IMPs in P-faces of normal^' muscle sarcolemmas ranging from 170/pm2 to 1,500/pm2 to 2800/pm2 (Rash, 1979; Schmalbruch, 1979; Schotland et al., 1977; Shafiq et al., 1979; Yoshioka and Okuda, 1977). Likewise, dystrophic myofibers were reported to exhibit from 110/pm2 (Schotland et al., 1977) to 1,000/pm2 (Schotland et al., 1980) to more than 3,000/pm2 (Rash et al., 1979b; ef. Shotton, 1982). By comparison to “leg counting” as a means of identifying animal species, this tenfold difference in IMPs in normal muscle and 30-fold difference in dystrophic myofibers is roughly the difference between the num-

Received October 20, 1988; accepted in revised form October 30, 1988. Address reprint requests to Dr. John Rash, Department of Anatomy and

Neurobiology, Colorado State University, Ft. Collins, CO 80523.

0 1989 ALAN R. LISS. INC.

CRITERIA FOR COUNTING AND MEASURING IMPs AND PITS 205

ber of legs on a cow and on a centipede. What is the basis for such large differences in IMP counts? Is it normal biological variability, or do these large differences reflect common but often unrecognized specimen preparation artifacts? Can such disparate data be used profitably to characterize any diseases? If not, is it possible to establish methods to eliminate such artifactual variability and to obtain reliable and consistent estimates of IMP density in normal tissues, which may then be used to identify changes in IMP profiles in disease? Finally, is it possible to establish minimum criteria that must be met before accepting IMP counts as valid or useful?

We have chosen rat and human skeletal myofibers as model systems because these cells allow several important artifacts that alter IMP counts to be demonstrated unambiguously and in a manner not possible with the membranes of other tissues. Thus, the purposes of this report are threefold: 1) to document two extremely common sources of specimen preparation artifact that routinely and dramatically alter IMP counts, 2) to demonstrate simple alterations in procedure that substantially increase the relative proportion of replica surface area that is useful for counting and measuring IMPs and pits, and 3) to establish criteria that will allow the researcher to determine which replicas are acceptable for obtaining IMP counts and measurements.

MATERIALS AND METHODS Sources of tissue

Normal rat extensor digitorum longus (EDL) mus- cles were obtained following perfusion fixation with 2.5% glutaraldehyde in well-oxygenated Ringer’s buffer (pH 7.4). For comparison with human myofibers, additional samples of rat EDL muscle were fixed by immersion in 2.5% glutaraldehyde and dissected so that surface and near surface fibers could be freeze- fractured. Details of methods are described in Rash and Ellisman (1974) and Rash (1983a,b). For analysis of human myofibers, intact (uncut) intercostal and biceps myofibers from normal subjects and patients with Duchenne muscular dystrophy were obtained by surgi- cal biopsy and fixed by immersion in 2.5% glutaraldehyde (cf. Rash et al., 1981).

Cryoprotection and freezing Samples were infiltrated with 30% glycerol in Ring-

er’s buffer and maintained at that final concentration for a t least 30 minutes before freezing. Samples were frozen by plunging into a slurry of freezing Freon 12 as described in Hudson et al. (1979,1981), except that the freezing solution was stirred with the blunt end of an EM forceps until the “freezing well” was almost completely refrozen. (This modification insures that cryogen is a t its freezing temperature when the sample is introduced.) The samples were plunge frozen, held at the liquid-frozen Freon interface for 5 seconds, and then “snapped into a liquid-nitrogen-filled storage Dewar flask. To estimate the relative frequency and extent of prefracturing resulting from uneven cooling rates throughout the sample, several samples were frozen in Freon 12 that had been allowed to adsorb and

condense water vapor Ke., from the atmosphere and from expelled breath). In this case, the liquid Freon was an opalescent slurry containing numerous visible ice crystals. These crystals serve as a “tracer” which infiltrates cracks and fissures created during freezing (Steere et al., 1980). Other samples were frozen in freshly condensed, clear (nonopalescent) Freon 12.

Comparing prefractured vs. knife-fractured membrane faces

To assess the propensity for muscle plasma membranes to undergo mechanical prefracture and subsequent internal “self-contamination” of exposed faces, samples were fractured with a fresh Shick Injector razor blade by using multiple knife passes (“shaved) or a single deep cleave (“karate cleaved”). Samples were replicated immediately after the final (or only) pass, as described by Ellisman and Staehelin (1979). Additional samples exhibiting deep cracks caused by internal freezing stresses (“cryogenic stress fractures”, cf. Steere et al., 1980) were evacuated to <1.5 x mBar and cleaved in a manner to widen and expose the visible fissures. Once exposed, these “cryogenic stress fracture” faces were replicated and the samples were thawed, carefully oriented, and processed to allow direct comparison with other areas of the same sample that were thought to have remained unfractured until cleaved by the knife.

Replication The three major groups of test samples were frac-

tured and replicated in a Balzers 301 freeze-etch device under “conventional” conditions (<1.5 x mBar pressure, 120 Hz platinum thickness deposited at an angle of 45” by using an electron beam gun operated at 2,200-2,500 V and 70-100 mA heating current). Plat- inum film thickness was measured and regulated by a quartz crystal film thickness monitor.’

Effects of deliberate contamination by water vapor

To assess the effects of multiple knife passes on the ability to recognize, count, and measure IMPs and pits following deposition of very thin layers of water vapor, additional rat and human myofibers were either “shaved” or “karate cleaved and subsequently exposed to vacuum conditions previously shown to produce layers of water vapor contamination of known thickness (<10A vs. 20 A and 40 A; cf. Rash et al., 1979a).

Electron microscopy Thin sections of undigested, plastic-embedded repli-

cas (cf. Rash, 1979) were photographed on a vintage

‘In a separate report (Rash et al., in preparation), we describe our efforts to improve the resolution of the platinum shadow by using much lower specimen tem eratures (-170°C) and substantially improved vacuum conditions (2 X

helium-cooled cryopump. (These near-“ultrahigh vacuums were measured by using a calibrated Balzers quadrupole mass spectrometer.) In those studies, an electron beam gun was tested over low to high voltage (1,600-2,500 V) at low to high heating currents (70-150 mA) and was used to deposit thin to thick platinum films (35 to 200 Hz). Except for an improved ability to resolve IMPs and pits less than 35 A ~n diameter, those data do not alter the major observations and conclusions described in this report.

10- 8 mBar H20 vapor pressure) obtained by using an 8“ CTI, Inc., liquid-

206 J.E. RASH AND F.D. GIDDINGS

Hitachi HS-8 TEM operated at 50 kV. For high- resolution analysis, replicas were examined at 80 or 100 kV on a Philips 400T TEM equipped with a k60” tilting stage (demonstrated resolution of 2.2 A obtained by using a gold lattice image) or at 80 kV on a Siemens 101 TEM equipped with a k24” double tilt device (4 A resolution).

Methods for counting and measuring IMPs and pits

Measurements and counts of IMPs and pits were made on near-focus and slightly underfocussed negatives and prints at 100,000~ and 150 ,000~ magnification by using a l o x Peak scale loupe. Stereo pairs were examined to insure that replicas were photographed at local viewing angles normal to the electron beam and devoid of appreciable contour (i.e., flat). Clear plastic overlays inscribed with an area equivalent to 0.1 pm2 were placed over each micrograph, and IMPs and pits were examined and measured indepen- dently by both authors. As each particle was counted and measured, it was obliterated on the overlay by using a permanent marker. After all IMPs were obliterated, pits were much more easily discerned. Pits were counted, measured, and marked by using a permanent marker of a different color. After each area was examined, the corresponding plastic overlays were superimposed to assess variabilityibias in identifying IMPs and pits. Discrepancies in counts of P-face IMPs vs. E-face pits (which initially were less than 5% in each of the areas examined) were reconciled by reex- amination of original micrographs by using stereo- scopic images. Final discrepancies in IMP counts from identical areas were less than 3%. (The freeze-fracture nomenclature of Branton et al., 1975, is used in this report.)

Areas occupied by caveolae and by “square arrays” (Kreutziger, 1968; Rash and Ellisman, 1974; Rash et al., 1973) were subtracted from the inscribed area and the resulting number of “dispersed’ IMPs was calcu- lated per pm2 of membrane P-face (cf. Shotton, 1982). Since “square arrays” contain an average of about 20 IMPs per array, and vary greatly in density (from O/pm2 near the endplate to over 30/pm2 at 0.5 mm from the endplate, and about 10/pm2 at >2 mm from the endplate, Ellisman et al., 19761, this may represent up to 500 square array IMPs/pm2 (Schotland et al., 1981). It is also relevant to note that square arrays provide precedent for local variations in IMP density as a function of distance from the neuromuscular junction as well as for variations in density as a function of fiber types (Ellisman et al., 1976; Rash and Ellisman, 1974). Unfortunately, it has not been ascertained whether the density of dispersed IMPs varies in a manner similar to the square arrays. However, this possibility cannot presently be excluded as a factor affecting IMP counts in various myofiber types and at unknown distances from the motor endplate.

Measurement of IMP and pit diameters were estimated to 5 A and presented as histograms reflecting number of particles and pits per pm2 at each 20 A-diameter increment from 20 A to 160 A. For detailed analyses of IMPs and pits, P- and E-face images from

within the same replica and separated by less than 100 pm were photographed. The E- and P-face images analyzed, therefore, are not precise complements but instead represent “complementary-type’’ images. We chose not to examine true complementary images because we have found that freezing rates are significantly slower in the relatively more massive 4 mm Balzers “double replica” planchettes (unpublished observations). The resulting increased freezing compres- sional forces produce substantial alterations in IMP cleaving patterns and morphologies. Hence, we analyzed only adjacent “complementary-type” images.

Particle diameters were measured at the widest margin within the area devoid of platinum, according to the convention of Staehelin (1976) (i.e., within the IMP platinum-free “shadow”). Likewise, pits were measured at the widest point within the area devoid of platinum (i.e., in the Pt-free region within the pit). This convention minimizes errors in estimating IMP diameters caused by variable-thickness platinum films. The resulting particle measurements from sarcolemmal P-faces and the pit measurements from the “complementary-type’’ E-faces are presented as mirror- image histograms. The thickness of all deposited layers (water vapor contamination plus thickness alteration by platinum coats) was then estimated by comparing the relative differences in diameter of the entire range of IMPs vs. that of the pits and dividing that difference by four (i.e., the added thickness to both sides of a particle plus the reduced diameter of pits caused by deposition on both sides of the pit).

RESULTS Pre-fracturing as a primary source of artifacts

in IMP counts Steere and co-workers (Steere et al., 1980) have

shown that conventional plunge freezing techniques yield numerous “cryogenic stress fractures” throughout the sample and that replicas of such prefracture faces, on occasion, may be recognized by the presence of cubic and hexagonal ice crystals. These crystal are thought to condense from the moisture in the surround- ing air and to crystalize in situ where the liquid cryogen penetrates the prefracture fissures. Compared with very-small-diameter yeast cells, mammalian skeletal muscle myofibers provide a more severe test of freeze-fracture methodologies. Foremost, the myofibers (individual muscle cells) are extremely long (many millimeters) and of very large diameter (50-100 pm). This provides for very large fracture planes that, once initiated, may inadvertently be extended entirely through the tissue, especially during multiple cleaves (“shaving”). Moreover, any source of mechanical or thermal stress a t any stage following initial entry of the sample into/onto the cryogen may cause “secondary” prefracture fissures to propagate partially or completely through the tissue. Sometimes these prefracture fissures are visible after the sample is mounted on the specimen stage for cleaving, but often they are not discernible.

At low magnification (Fig. lA), cryogenic stress fracture faces from muscle cells frozen in Freon 12 saturated with frozen water vapor are often demar-

CRITERIA FOR COUNTING AND MEASURING IMPS AND PITS 207

Fig. 1. Rat muscle frozen in Freon 12 saturated with water vapor. After fracturing and replication, hexagonal and cubic ice crystals provide a structural tracer that delineates the depth of cryogenic prefracture fissures (CPF, Cf. Successively higher magnifications depict the interface between cryogenic fracture faces and knife- fracture faces (CPFIKFF, C). IMPS and pits are virtually obliterated

by wateriice contamination deposited either during specimen storage or during subsequent evacuation steps. (The area labeled KFF? was exposed during knife cleaving but may represent a “secondary” mechanical fracture that could have occurred at any time subsequent to initial freezing.) A, x 11,500. B, x 26,500. C, x 100,000.


cated by the presence of replicated cubic and hexagonal ice crystals both on cleaved membranes and on cleaved cytoplasm (Fig. lA, asterisks). At higher magnification (Fig. lB,C), IMPs and pits on both sides of the margin of the prefracture may be compared. In the area that apparently remained unfractured until cleaved by the knife (i.e,, beyond the prefracture margin), IMPs are distinct and well resolved, both within the “square arrays” and dispersed in the muscle plasma membrane. In the cryogen-induced prefracture faces (CPF), IMPs exhibit one of two distinct morphologies: 1) Near the deepest extent of the prefracture fissure, IMPs are less distinct. They appear partially crushed andior covered by a layer of contamination (Fig. 1C). 2) On the remainder of the cryogen-induced fracture face, however, IMPs at first glance appear relatively uncontam- inated by water vapor, even in those areas exhibiting 0.25-0.75 pm cubic and hexagonal ice crystals. Pre- sumably, at least some etching (vacuum sublimation) occurred after the specimen had warmed to -105°C and prior to conventional (intentional) knife fracture. Thus, the relative amount of frost removed during evacuation was dependent on a) the local cross- sectional area of the narrowing V-shaped fissure (unknown, but the crack contained 0.75 pm ice crystals), 2) the distance to the exterior opening of the fissure (unknown), 3) the temperature of the specimen at the final cleave (- l05”C), and 4) the time of evacuation/ etching (unknown, but etching was presumed to occur only after the sample temperature was warmed to above - 120”C, the approximate “devitrification” temperature of 30% glycerol ([Luyet and Kroener, 1966; Luyet et al., 1967; see also Franks, 1973; Rash, 198313; Steinbrecht and Muller, 1986; Washburn, 19241). Be- cause of the variability in etchinglcontamination rates inherent to such prefractures, we have made no at- tempt to quantify IMPs on cryogenic stress fracture faces.I2

Mechanical prefracturing during conventional knife fracturing (“shaving” vs.

“karate cleaving”) Muscle cells, like all other eukaryotic cells, contain

numerous membrane components in which fracture planes may be initiated. Once initiated in nuclear, mitochondrial, sarcoplasmic reticulum, other membranes, or even in aqueous cytoplasm and nucleoplasm, those “secondary” stress fracture planes will be extended for variable distances away from the primary fracture plane, at least until mechanical resistance of the tissue and/or decreased local application of mechanical force leads to abortive termination of the fracture plane (Fig. 2A,B). Successive mechanical or thermal stress may 1) extend these secondary fissures past the debris margin that is sometimes observed,

‘Similarly, ultrarapid freezing devices as well as subsequent “shaving” may also yield mechanical prefractures, but unless the samples are stored in liquid cryogen containing dissolved water molecules, they may not exhibit telltale ice crystals (see next sections). Water vapor contamination of prefractured surfaces in rapidly frozen myofibers may, nevertheless, occur during sample transfers, during evacuation of the vacuum chamber, or preceding each knife pass (see Fig. 6 in Rash et al., 1979a). The extent of “secondary” prefracturing and subsequent self-contamination of closely apposed surfaces is described below.

2) initiate additional fissures, and/or 3) expose for replication membrane faces created by previous secondary fractures (Fig. 2B,C).

To document the relative amount of secondary fracturing that routinely occurs during each pass of the knife, a single moderate-depth pass was made by passing the knife slowly through an apparently solid, well frozen, cryoprotected sample of rat muscle. The sample was immediately shadowed at, 45” with platinum, rotary replicated with carbon at an angle of 75”, thawed, postfixed with Os04, dehydrated, embedded in plastic, and sectioned parallel to the plane of knife fracture. Before and after embedding, the tissue ap- peared to be solid and undamaged. However, a t depths of up to 50 pm beneath the knife fracture plane, numerous cracks and fissures well below the limits of resolution of the binocular viewer were observed in thin sections (Fig. 2C). Occasionally, the fissures were oriented so that they were penetrated and delineated by the Pt/C replica (Fig. 2C, arrowheads); some were delineated by carbon only, and others were not penetrated by Pt or C. The magnitude and widespread nature of the secondary stress fractures arising from a single “karate cleave” may be discerned in the fragments of myofibrils that were displaced and/or rotated (Fig. 2C, asterisks) as well as in the dozens of additional cracks (arrows) that permeate the sample. Un- fortunately, the closely apposed surfaces within the fissures provide both source and sink for water vapor, providing the potential for the deposition of variable thickness layers of snow^' or “frost” as “self-contamination” (Fig. 2A,B, large arrows). These contaminating layers of condensed water vapor may alter or obliterate replica detail. It must be also be assumed that the secondary stress fracture faces become “self-contaminated” whenever the sample temperature is raised above the devitrification temperature (approximately -120°C). The high frequency of stress fractures in freeze-fractured blocks that we have examined by thin sectioning suggests the likelihood that similar sources of contamination have occurred frequently in all previous freeze-fracture studies of muscle plasma membranes fractured by “shaving” methods. Without alter- native clues such as detailed particle counting and measuring to ascertain relative complementarity of membrane faces, these “secondary prefracture” fissures would be difficult to identify (see Fig. 3D). Criteria for identifying secondary mechanical fracture faces and discriminating them from knife fracture faces (KFF) are presented below.

Relative proportion of prefractured vs. knife-fractured areas

In detailed analyses of conventional replicas, prefractured areas have been found to be much more extensive than anticipated, often exceeding more than 75% of the replicated fracture face (Fig. 3A,B). More- over, when replicas fragment, areas recovered for TEM examination may consist entirely of prefractured surfaces and faces. Thus, the use of morphometric techniques requiring “random” selection of areas to be analyzed will necessarily yield inaccurate estimates of IMPs, regardIess of the number of images analyzed.


Fig. 2. Diagrams and corresponding thin section image illustrating the extent of mechanical fracturing that occurs during conventional knife cleaving. Rather than conventional cleaning, this typical sample was embedded and sectioned parallel to the original knife- fracture plane. The PtiC replica (arrowheads) penetrated cracks that were oriented toward the electron beam gun. Other cracks (arrows) do

not contain either platinum or carbon. If the sample is cleaved at a temperature above approximately - 120°C, the closely apposed surfaces of prefracture fissures provide both source and "sink for water vapor contamination. If subsequently exposed by successive knife passes (B), these faces are likely to be contaminated and of no value for IMP counts. C, x 6,000.


Fig. 3. Rat muscle frozen in freshly condensed Freon 12 (containing little condensed water vapor). After fracturing and replication, there are no large telltale hexagonal or cubic ice crystals to provide a tracer for delineating cryogenic prefracture fissures. Nev- ertheless, successively higher magnifications reveal that more than 75% of this replica had been prefractured (A, the area to the right of the apposed arrowheads). A faint linear deposit of debris (frozen,

noncrystalline water?) marks the deepest extent of the prefracture fissure (Cj. Not coincidentally, the debris marks the edge of the region of higher electron density of the platinum replica. This change in opacity provides a second marker for differentiating between cryogen prefractured faces vs. knife-fractured faces. (The inscribed rectangle in each micrograph is enlarged as the succeeding micrograph.) A, x 740, B, x 4,000, C, x 60,000. D, x 111,000.


Proper selection techniques will require a preliminary assessment of replica quality in the immediate vicinity from which the images will be obtained for morphometric analysis. The criteria used by Schotland et al. (1980) represent substantial progress toward that goal. Nevertheless, additional factors must now be consid- ered.

In samples that were not frozen or stored in liquid cryogen containing ice crystals, distinctive hexagonal and cubic crystals are usually not observed on membrane fracture faces. Nevertheless, it is still possible to identify the demarcation between cryogenic prefracture areas (CPF) vs. other areas thought to represent membranes exposed by knife fracture (KFF). The deepest extent of the primary pre-fracture fissure may be recognized by either of two distinct features: 1) a faint linear deposit of unidentified debris (deposits of dissolved frost?) that infiltrated the fissure during specimen immersion in the cryogen (Fig. 3C, arrow with asterisk, designating the “tide mark”), and 2) an as yet incompletely characterized artifact of platinum shadowing in which an increased electron density was observed on the contaminated CPF. Examination of negatives and prints at very high magnification suggests that this increased electron density reflects a more nearly confluent distribution of platinum grains on CPF surfaces. The thin layer of contamination deposited before knife fracture results in a smoother microcontour. As a result, contaminated replicas have lower relief, shorter shadows, and a lower proportion of the replica devoid of platinum, thereby decreasing electron transmittance. (In addition, platinum grains may have different states of nucleation when deposited on a thin layer of ice vs. on freshly exposed membrane faces, possibly contributing to the observed difference in electron opacity.) In the absence of “high tide marks” or abrupt changes in electron density, only laborious particle and pit counting and measuring techniques can be used to establish relative complementarity (see below), as aids in determining whether membrane faces are obtained from cryogen-induced fractures, stress prefractures, or knife fractures.

“Relative complementarity”: the effects of controlled contamination on number and

diameter of IMPS and pits

Chalcroft and Bullivant (1970) and Steere (1971) proposed that the essential criterion for assessing replica quality was the relative “complementarity of fit” of membrane fracture faces. In fracturing rigid (frozen) samples, unless mass is created, destroyed, or altered in shape (“plastic deformation”), the two (or more) fragments of the original sample must be capable of fitting back together. In the case of membrane split- ting, this means that every IMP on one face must have a corresponding pit of equivalent size on the “complementary” fracture face (Steere, 1971; Ting-Beall et al., 1986). It was also recognized that there would be one primary exception to this rule-that of “plastic deformation” (Bohler, 1979; Sleytr and Robards, 19771, in which membrane integral proteins would be stretched or deformed during the fracturing process and (by

implication) not return to their original shapes (i.e., not undergo “elastic rebound). Images of plastic deformation reveal unmatched features in the two otherwise complementary fracture faces (e.g., “strings” opposite conventional pits; cf. Steere et al., 1980). In succeeding years, many exceptions to the “rule of complementarity” have been proposed based on the apparent absence of pits opposite well-defined IMPs. In the muscle sarcoplasmic reticulum, for example, many reports have described a “rough” or particulate P-face and a “smooth” E-face devoid of pits (Bertaud et al., 1970; Bray and Rayns, 1976; Deamer and Baskin, 1969; Rash et al., 1974). However, improvements in shadowing procedures permitting virtually immediate shadowing of freshly exposed fracture faces (Ellisman and Staehe- lin, 1979) have allowed the requisite array of pits in the E-face to be demonstrated (Rash et al., 1979a; Sommer, 1980). The source of this artifactual noncomplementarity was traced to extremely rapid and preferential deposition of water vapor on the hydrophilic transmembrane pits.

To assess the effects on complementarity of muscle plasma membranes caused by minute layers of water vapor contamination deposited on prefractured faces, we have utilized methods that allow semiquantitative deposition of water vapor at established rates (Rash et al., 1979a; Steere et al., 1979). Following a single “karate cleave” and immediate replication (Ellisman and Staehelin, 19791, plasma membranes of Duchenne myofibers exhibit P-face IMPS at a density of 3S91/pm2 (Fig. 4A, and as a histogram from that same image in Fig. 5A). Similarly, E-face pits are present in the adjacent “complementary-type’’ images at a density of 3214/pm2 (Figs. 4D vs. 5A; standard deviations are not presented because these data represent single measurements from the individual images shown). De- tailed measurements reveal that IMP and pit diameters are very similar, with peaks apparently offset by approximately 20 A. Since the growth of IMPs and shrinkage of pits is based on deposition of water vapor on opposite sides of both, we presume that the 20 A difference in IMP vs. pit diameters reflects an average layer of contamination of less than 5 A thickness. Thus, these replicas represent near precise complementarity of P- and E-faces.

In contrast, deliberate contamination of freshly fractured surfaces b layers of approximately 20 A (Fig. 4B

in number of both IMPs and pits to 1,7841pm2 VS. 742/pm2 and 876/pm2 vs. 308/pm2, respectively. IMPS in the square arrays remain, but the overall textures of the membranes are consistent with increasing thickness blankets of “frost” or “snow.” In comparison, pits were preferentially depleted, but they were never completely obliterated, even with >40 A of water vapor contamination (Figs. 4C,F, 5C). The decreasing numbers of IMPs with increasing water vapor contamination are approximately the same as those reported by other investigators as reflecting changes attributable to disease etiology (references cited above). However, based on a firm acceptance of the rule of complementarity, we believe that the higher and relatively identical numbers for P-face IMPs vs. E-face pits are more

vs. 4E) and 40 Al (Fig. 4C vs. 4F) resulted in reduction


DUCHENNE SARCOLEMMA RAT SARCOLEMMA I CIO H Contamination I 20 H Contamination

RAT SARCOLEMMA 40 Contomination

876 IMP'S

m2

20 40 60 80 100 120 140 160

308 Pits


Fig. 6. Diagram illustrating to scale the progressive obliteration of surface detail by 5 A, 20 A, and 40 A of water vapor Contamination. Compare to images of IMPs and pits in Figure 4A-F.

accurate. We interpret the lower and noncomplemen- tary values as follows: With increasing thicknesses of water vapor contamination, the smaller IMPs and pits are progressively obliterated by a blanket of frost (Fig. 6A-C), while the pits are progressively filled in, either by specific “decoration artifact” (Gross, 1979) of the hydrophilic transmembrane channel or simply by frost

Fig. 4. Paired P-face and E-face replicas of muscle plasma membranes. Each pair has successively greater thicknesses of water vapor contamination (5 A in A, D 20 b in B, E; and 40 A in C, F). Pits are indicated by arrowheads. Particles and pits progressively disappear as the growing blanket of frost successively obliterates membrane detail. F, x 150,000.

Histograms of IMPs on membrane P-faces vs. pits on membrane E-faces. All histograms represent counts and measurements from the individual images in Figures 4A-F. Near precise complementarity is documented when the layer of water vapor contamination is less than 5 A thick.

Fig. 5.

blanketing. Nevertheless, we emphasize that IMPs and pits of all diameters were observed after deposition of layers thicker than 40 A. Thus, residual small-diameter pits may represent partially filled pits that were originally of larger diameter. Alternatively, they may represent new “etch pits” in the confluent snowy blanket. The drastic reduction in IMP numbers ap- pears to result from closely spaced small-diameter IMPs being converted first to single large-diameter IMPs (Fig. 6B), which, with increasing contamination, ultimately disappear beneath the blanket of frost and snow (Fig. 6C). Additional “pseudoparticles” represent- ing forming cubic and hexagonal ice crystals (Gross, 1979) may also be counted as IMPs, thereby accounting for “IMPs” even on grossly contaminated membrane faces (Figs. 4C, 6C).

Very few IMPs or pits larger than 120 A diameter were observed in replicas with contaminating layers of water vapor less than 20 A thick. In contrast, published reports show images with numerous IMPs larger than 140 A and 200 A, respectively (Ishikawa et al., 1975; Schotland, 1977). These very-large-diameter IMPs have been traced to localized leaks in the vacuum chamber (Steere et al., 1979). Consequently, we note that although “deep etching” might be useful in lower- ing the absorbed layer of ice to reveal IMPs and pits, inadequate etching may reveal variable numbers of IMPs and pits protruding from an incompletely etched layer of contamination. On the other hand, attempting further etching by using a machine with unidentified leaks in the vacuum chamber would contribute to further formation of frost (Steere et al., 1979). In the presence of unknown sources of water vapor, the likelihood of nonreproducible exposure of IMPs by at- tempts a t deep etching would require extra caution in the interpretation of IMP/pit profiles.

Finally, we note that of several thousand replicas that we have obtained following conventional fracturing (i.e., “shaving”), our IMP counts from sarcolemmal P-faces ranged from 1600/~m’ to 2,400/pm2, but in each case with evident noncomplementarity (i.e., with substantially fewer E-face pits than P-face IMPs; Rash, unpublished observations). In contrast, our more recent “karate cleaves” have yielded replicas with 3,200 P-face IMPs per pm2 and an equal number of E-face pits. Since the lower values (1,600-2,400/pm2) are similar to or greater than published values from skeletal muscle, we conclude that mechanical and/or cryogenic prefracturing has been an important but often overlooked factor in altering IMP counts in most published studies. Moreover, since this variable may produce substantial alteration of IMP counts independent of and of greater magnitude than the changes reported in various disease etiologies, we conclude that all published studies of IMP counts should be examined critically.

INTERPRETATION In this report, we have compared the number of

P-face IMPs with the number of E-face pits in a complex, naturally occurring membrane. By identifying prefractured vs. knife-fractured membrane faces, we have been able to identify areas exhibiting near


precise complementarity of fit. However, we explicitly acknowledge that very-small-diameter IMPs and pits may have been undetected in the background granu- larity of platinum films deposited from electron guns (ca. 25 A, cf. Zingsheim et al., 1970). Thus, i t is ex- pected that finer grain films (especially those obtained by using thinner deposits of metal or different shadowing materials, cf. Bridgman et al., 1988) will yield replicas with higher resolution and higher IMP counts, especially with respect to IMPs and pits smaller than 30 A. For example, we recently demonstrated the existence and subsequent “patching” of one class of very-small-diameter IMPs corresponding to membrane immunogIobins in mouse B lymphocytes (Dinchuk et al., 1987). The transmembrane portion of the mIg molecule is thought to be a single a-helix composed of approximately 22 amino acid residues, thereby repre- senting the smallest-possible-diameter IMP. Those small IMPs were so near the limit of resolution of current replication techniques that nonpatched mIg molecules were not definitively identifiable in our replicas; nor were they resolved in previous studies of patched mIg (Karnovsky and Unanue, 1972). Only following patching and immunogold labeling could they be discerned with confidence (Dinchuk et al., 1987). Based on the difficulty in resolving such small- diameter IMPs, we do not propose that the replicas described in this report reflect “true” numbers of IMPs. Rather, these images were presented as a way of illustrating minimal criteria for recognizing and elimi- nating inadequate or misleading freeze-fracture images before IMP counts are made or inappropriate conclusions are drawn from those counts.

When techniques are adopted in which individual classes of IMPs are reproducibly replicated from laboratory to laboratory, it may become feasible to assign functional identity to individual IMPs. If this presump- tion is valid, the development of “label-fracture,” “fracture-label,” “shadow-label,’’ and “replica-label” techniques (Dinchuk et al., 1987; Pinto da Silva and Kan, 1984; Pinto da Silva et al., 1981a,b; Rash et al., 1978, 1982; reviewed by Rash et al., 1989; and Severs, this issue) could transform freeze-fracture from a purely morphological and descriptive discipline into a powerful analytical tool of physiology, pharmacology, and biochemistry. For example, with the isolation of dystrophin, the membrane protein that is deficient or altered in several of the muscular dystrophies (Hoff- man et al., 19881, the production of antidystrophin antibodies (Hoffman et al., 1987a,b), and identification of dystrophin as a membrane protein of the T-tubules and sarcolemma (Bonilla et al., 1988; Hoffman et al., 198713; Sugita et al., 19871, molecular localization within muscle plasma membranes may be feasible by replica-labeling techniques. Demonstration of near precise complementarity of membrane faces is an essential criterion for obtaining that goal.

ACKNOWLEDGMENTS This work was performed at the University of Mary-

land School of Medicine, Baltimore, MD, from 1978 to 1979 and a t Colorado State University, from 1979 to 1988 under funding provided by NIH (NS 14648 and

NS 15991) and from the MDA. This report was pre- pared with funds provided by BRSG.

REFERENCES Bertaud, W.S., Rayns, D.G., and Simpson, F.O. (19701 Freeze-etch

studies on fish skeletal muscle. J. Cell Sci., 6:537-557. Bohler, S. (1979) Artifacts and defects of preparation in freeze-etch

technique. In: Freeze-Fracture: Methods, Artifacts, and Interpreta- tions. J.E. Rash and C.S. Hudson, eds. Raven Press, New York, pp. 19-29.

Bonilla, E., Fishbeck, K., and Schotland, D.L. (1981) Freeze-fracture studies of muscle cavcolae in human muscular dystrophy. Am. J . Pathol., 104:167-173.

Bonilla, E., Sammit, C.E., Miranda, A.F., Hays, A.P., Salviati, G., Dinauro. S.. Kunkel. L.M.. Hoffman, E.P., and Rowland, L.P. (1988) Duchenne muscular dystrophy deficiency of dystrophin at the muscle cell surface. Cell, 54:447-452.

Branton, D., Bullivant, S., Gilula, N.B., Karnovsky, M.J., Moor, H., Muhlethaler, K., Northcote, D.H., Packer, L., Satir, B., Satir, P., Speth, V., Staehelin, L.A., Steere, R.L., and Weinstein, R.S. (1975) Freeze-etching nomenclature. Science, 190:54-56.

Bray, D.F., and Rayns, D.G. (1976) A comparative freeze-etch study of the sarcoplasmic reticulum of avian fast and slow muslce fibers. J . Ultrastruct. Res., 57:251-259.

Bridgman, P.C., Carr, C., Pedersen, S.E., and Cohen, J.B. (1988) Visualization of the cytoplasmic surface of Torpedo postsynaptic membranes bv freeze-fracture and immunoelectron microscouv. J.

A “

Cell Biol., 10~:1829-1854. Chalcroft, P., and Bullivant, S. (1970) An interpretation of liver cell

membrane and junction structure based on observation of freeze- fracture replicas of both sides of the fracture. J. Cell Biol., 47:49- 60.

Deamer, D., and Baskin, R.J. (1969) Ultrastructure of sarcoplasmic reticulum preparations. J . Cell Biol., 42:296-307.

Dinchuk, J.E., Johnson, T.J.A., and Rash, J.E. (1987) Post-replication labeling of E-leaflet molecules: Membrane immunoglobulins localized in sectioned labeled replicas examined by TEM and HVEM. J . Electron Microsc Tech., 7:l-16.

Ellisman, M.H., Rash, J.E., Staehelin, L.A., and Porter, K.R. (1976) Studies of excitable membranes. 11. A comparison of specializations at neuromuscular junctions and non-junctional sarcolemmas of mammalian fast and slow twitch muscle fibers. J. Cell Biol., 68:752-774.

Ellisman, M.H., and Staehelin, L.A. (1979) Electronically interlocked electron gun shutter for preparing improved replicas of freeze- fracture specimens. In: Freeze-Fracture: Methods, Artifacts, and Interpretations, J.E. Rash and C.S. Hudson, eds. Raven Press, New York, pp. 123-125.

Franks, F. (1973) Water, a comorehensive treatise. In: Water in Crystalline Hydrates: Aqueous- Solutions of Simple Nonelectro- lytes. Plenum, New York. Vol. 2 pp. 376-399.

Gross, H. (1979) Advances in ultrahigh vacuum freeze fracturing a t very low specimen temperature. In: Freeze-Fracture: Methods, Artifacts, and Interpretations. J.E. Rash and C.S. Hudson, eds. Raven Press, New York, pp. 121-139.

Hoffman, E.P., Brown, B.H., Jr., and Kunkel, L.M. (1987a) Dystro- phin: The protein product of the Duchenne muscular dystrophy gene. Cell, 51:919-928.

Hoffman, E.P., Knudson, C.M., Campbell, K.P., and Kunkel, L.M. (198713) Subcellular fractionation of dystrophin to the triads of skeletal muscle. Nature, 330:754-758.

Hoffman, E.P., Fischbeck, K.H., Brown, R.H., Johnson, M., Medori, R., Loike, J.D., Harris, J.B., Waterston, R., Brooke, M., Specht, L., Kupsky, W., Chamberlain, J . , Caskey, C.T., Shapiro, F., and Kunkel, L. (1988) Characterization of dystrophin in muscle-biopsy specimens from patients with Duchenne’s or Becker’s muscular dystrophy. N. Engl. J. Med., 318:1363-1368.

Hudson, C.S., Rash, J.E., and Graham, W.F. (1979) Introduction to samde ureuaration for freeze-fracture. In: Freeze-Fracture: Meth- ods, Artifac‘ts, and Interpretations. J.E. Rash and C.S. Hudson, eds. Raven Press, New York, pp. 1-10.

Hudson, C.S., Rash, J.E., and Shinowara, N.L. (1981) Freeze-fracture and freeze-etch methods. In: Current Trends in Morphologcal Techniques. Vol. 11. J . J . Johnson, ed. CRC Press, Inc.. Boca Raton. pp. 183-217.

Ishikawa, H., Fukuda, Y., and Yamada, E. (1975) Freeze-replica observations on frog sartorius muscle. I. Sarcolemmal specializations. J. Electron Microsc. (Tokyo), 24:97-107.


Karnovsky, M.J., and Unanue, E.R. (1972) Mapping and migration of lymphocyte surface macromolecules. Fed. Proc., 3255-59.

Ketelsen, U.-P. (1975) The plasma membrane of human skeletal muscle cells in the pathological state. Freeze-etch studies. In: Recent Adv. Myol., Excerpta Med. (Amsterdam) Int. Cong. Ser.,

Kreutziger, G.O. (1968) Freeze-etching of intercellular junctions of mouse liver. In: Proc. 26th Meet. EMSA. G.W. Bailey, ed. Claitors Publishing Div., Baton Rouge, p. 234.

Luyet, B.J., Sager, D., and Gehenio, P.M. (1967) The phenomenon of “premelting recrystallization”. Biodynamica, 10:123-132.

Luyet, B.J., and Kroener, C. (1966) The temperature of the “glass transition” in aqueous solutions of glycerol and ethylene glycol. Biodynamica, 10:33-40.

Osame, M., Engel, A.G., Rebouche, C.J., and Scott, R.E. (1981) Freeze-fracture electron microscopic analysis of plasma membranes of cultured muscle cells in Duchenne dystrophy. Neurology, 31:972- 979.

Pinto da Silva, P., Kachar, B., Torrisi, M.R., Brown, C., and Parkison, C. (1981a) Freeze-fracture cytochemistry: Replicas of critical point- dried cells and tissues after fracture-label. Science, 213:230-233.

Pinto da Silva, P., and Kan, F.K. (1984) Label-fracture: A method for high resolution labeling of cell surfaces. J . Cell Biol., 99:1156-1161.

Pinto da Silva, P., Parkison, C.P., and Dwyer, N. (1981b) Fracture- label: Cytochemistry of freeze-fracture faces in the erythrocyte membrane. Proc. Natl. Acad. Sci. USA, 78:343-347.

Rash, J.E. (1979) The sectioned replica technique: Direct correlation of freeze-fracture replicas and conventional thin-section images. In: Freeze-Fracture: Methods, Artifacts, and Interpretations. J.E. Rash and C.S. Hudson, eds. Raven Press, New York, pp. 153-160.

Rash, J.E. (1983a) Ultrastructure of normal and myasthenic end- plates In: Myasthenia Gravis. E.X. Albuquerque and M.E. Elde- frawi, eds. Chapman and Hall, New York, pp. 391-421.

Rash, J.E. (1983b) The rapid-freeze technique in neurobiology. Trends Neurosci., 6:208-212.

Rash, J.E., and Ellisman, M.H. (1974) Studies of excitable membranes. I. Macromolecular specializations of the neuromuscular junction and the nonjunctional sarcolemma. J . Cell Biol., 63567- 586.

Rash, J.E., Ellisman, M.H., Staehelin, L.A., and Porter, K.R. (1974) Molecular specializations of excitable membranes in normal dener- vated and dystrophic muscle fibers. In: Exploratory Concepts in Muscular Dystrophy (11). L.P. Rowland, ed. Excerpta Med. (Amster- dam) Int. Cong. Ser., 333:271-289.

Rash, J.E., Graham, W.F., and Hudson, C.S. (1979a) Sources and rates of contamination in a conventional Balzers freeze-etch device. In: Freeze-Fracture: Methods, Artifacts, and Interpretations. J.E. Rash and C.S. Hudson, eds. Raven Press, New York, pp. 111-122.

Rash, J.E., Hudson, C.S., and Ellisman, M.H. (1978) Ultrastructure of acetylcholine receptors at mammalian neuromuscular junctions. In: Cell Membrane Receptors for Drugs and Hormones: A Multidisci- plinary Approach. R.W. Straub and L. Bolis, eds. Raven Press, New York, pp. 47-68.

Rash, J.E., Hudson, C.S., Graham, W.F., Mayer, R.F., Warnick, J.E., and Albuquerque, E.X. (1981) Freeze-fracture studies of human neuromuscular junctions. Membrane alterations observed in myasthenia gravis. Lab. Invest., 445194531.

Rash, J.E., Hudson, C.S., Graham, W.F., and Mayer, R.F. (1979b) Freeze-fracture studies of patients wtih neuromuscular disease. Neurology, 29594-595.

Rash, J.E., Johnson, T.J.A., Hudson, C.S., Giddings, F.D., Graham, W.F., and Eldefrawi, M.E. (1982) Labelled-replica techniques: Post- shadow labelling of intramembrane particles in freeze-fracture replicas. J. Microsc., 128:121-138.

Rash, J.E., Johnson, T.J.A., Dinchuk, J.E., Duch, D.S., and Levinson, S.R. (1989) Labelling intramembrane particles in freeze-fracture

360:446-454.

replicas. In: Freeze-Fracture Studies of Membrane Structure. S.-W. Hui, ed. CRC Press, Boca Raton, in press.

Rash, J.E., Lee, R.E., and Walrond, J.E. (in preparation) Freeze- fracture Replication at lo-’’ mBar in a Balzers 301 Device Equipped With an 8 CTI Cryopump.

Rash, J.E., Staehelin, L.A., and Ellisman, M.H. (1973) Rectangular arrays of particles are not gap junctions. Exp. Cell Res. 86:187-190.

Schmalbruch, H. (1979) “Square arrays” in the sarcolemma of human skeletal muscle fibers. Nature, 281:145-146.

Schotland, D.L. (1977) Duchenne dystrophy-A freeze-fracture study. In: Pathogenesis of Human Muscular Dystrophies. L.P. Rowland, ed. Excerpta Med. (Amsterdam) Inter. Cong. Ser., 404562468.

Schotland, D.L., Bonilla, E., and Van Meter, M. (1977) Duchenne dystrophy: Alteration in muscle plasma membrane structure. Sci- ence, 196: 1005-1007.

Schotland, D.L., Bonilla, E., and Wakayama, Y. (1980) Application of the freeze-fracture technique to the study of human neuromuscular disease. Muscle Nerve, 3:21-27.

Schotland, D.L., Bonilla, E., and Wakayama, Y. (1981) Freeze- fracture studies of muscle plasma membrane in human muscular dystrophy. Acta Neuropathol. (Berl.), 54:189-197.

Shafiq, S.A., Leung, B., and Shutta, H.S. (1979) A freeze-fracture study of fiber types in normal human muscle. J . Neurol. Sci., 42:129-138.

Shotton, D.M. (1982) Quantitative freeze-fracture electron microscopy of dystrophic muscle membranes. J . Neurol. Sci., 57:161-190.

Singer, S.J., and Nicolson, G.L. (1972) The fluid mosaic model of the structure of cell membranes. Science, 175720-731.

Sommer, J . (1980) Freeze-fracture appearance of rabbit cardiac sarcoplasmic reticulum. In: Proc. 28th Annu. Meet. EMSA. Claitors Publishing Div., Baton Rouge, pp. 630-631.

Sleytr, U.B., and Robards, A.W. (1977) Plastic deformation during freeze-cleaving: A review. J . Microsc., 11O:l-25.

Staehelin, L.A. (1976) Reversible particle movement associated with unstacking and restacking of chloroplast membranes in uitro. J . Cell Biol., 71:136-158.

Steinbrecht, R.A., and Muller, M. (1986) Freeze substitution and freeze-drying. In: Cryotechniques in Biological Electron Micros- copy. Springer-Verlag, New York, pp. 149-172.

Steere, R.L. (1971) High resolution stereography of complementary freeze-fracture replicas reveals the presence of depressions opposite membrane associated particles of split membranes. Proc. 29th Annu. Meet. EMSA. G.W. Bailey, ed. Claitors Publishing Div., Baton Rouge, pp. 442-443.

Steere, R.L., Erbe, E.F., and Moseley, J.M. (1979) Controlled contamination of freeze-fractured specimens. In: Freeze-Fracture: Meth- ods, Artifacts, and Interpretations. J.E. Rash and C.S. Hudson, eds., Raven Press, New York, pp. 99-109.

Steere, R.L., Erbe, E.F., and Moseley, J.M. (1980) Pre-fracture and cold-fracture images of yeast plasma membranes. J . Cell Biol., 86113-122.

Sugita, H., Arahata, K., Ishiguro, T., Suhara, Y., Tsukahara, T., Ihiura, S., Eguchi, C., Nonaka, I., and Ogawa, E. (1987) Negative immunostaining of Duchenne muscular dystrophy and mdx muscle surface membrane with antibody against synthetic peptide fragment predicted from DMD cDNA. Proc. Jpn. Acad. [B[, 64:37-39.

Ting-Beall, H.P., Burgess, F., and Robertson, J.D. (1986) Particles and pits matched in native membranes. J . Microsc., 142:311-316.

Washburn, E.W. (1924) The vapour pressure of ice and of water below the freezing point. Monthly Weather Rev., 1924, pp 488-490.

Yoshioka, M., and Okuda, R. (1977) Human skeletal muscle fibers in normal and pathological states: Freeze-etch replica observations. J . Electron Microsc., 26: 103 -110.

Zingsheim, H.P., Abermann, R., and Bachmann, L. (1970) Shadow casting and heat damage. In: Proc. 7th Int. Cong. Electron Microsc. Grenoble, 1:411.

counting and measuring imps and pits: why accurate counts are exceedingly rare · 2008-11-04 ·...

Documents