| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Human Biology and Nutritional Sciences, University of Guelph, Guelph, N1G 2W1; 2 Department of Pathology and Molecular Medicine, and 3 Departments of Medicine and Kinesiology, McMaster University, Hamilton, L8S 4L8; and 4 Department of Kinesiology, University of Windsor, Windsor, Ontario, Canada N9B 3P4
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
A few qualitative investigations
suggested that location of muscle glycogen (G) granules in specific
sites may be associated with distinct metabolic roles. Similarly, it
has been suggested that the acid-soluble and -insoluble G fractions
(macro- and proglycogen, respectively) are different metabolic pools
and also could exist as separate entities. We employed a transmission
electron microscopic technique to quantify subcellular G particle size,
number, and location in human vastus lateralis biopsies of 11 resting
men. The intra- and interobserver variability for the various measures was generally <4%. Granule size and number were quantified in subcellular compartments (subsarcolemmal, intra- and
intermyofibrillar). Subcellular location was critical: G was more
densely concentrated in the subsarcolemmal than in the myofibrillar
space, whereas the single-particle volume was greater in the latter.
Single-particle diameter ranged from 10 to 44 
m and followed a
continuous, normal distribution. This implies that proglycogen is not a
distinct entity, but rather that pro- and macroglycogen are divisions
of smaller and larger molecules. These results demonstrate a
compartmentalized pattern of subcellular G deposition in human skeletal
muscle for both the size and density of granules.
glycosome; metabolic compartments; electron microscopy; carbohydrate; glycogen regulation; proglycogen; macroglycogen
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
SINCE THE DISCOVERY OF
GLYCOGEN by Claude Bernard, numerous investigators have addressed
many aspects of its metabolism. Although muscle glycogen concentration
has been routinely quantified biochemically, the subcellular
organization of glycogen particles has been studied much less
frequently and only with qualitative, descriptive transmission electron
microscopy (TEM) methods. Wanson and Drochmans
(31) performed the first comprehensive description
of rabbit skeletal muscle glycogen in its particulate 
-form.
Drochmans (5) had previously examined negatively stained
liver glycogen by using TEM and described three glycogen structures in
liver: 
-, -, and 
-particles. The -particles were the
typical liver rosettes, and the -particles were the 20- to 30-m spheroid units forming the -rosettes. The -particles
were identified as 3-m subunits of both - and -structures. The
single -particles described in muscles by Wanson and Drochmans
(31) corresponded in size and shape to the -subunits
that constituted the -rosettes in liver.
Scott and Still (28) proposed that particulate glycogen
was not a molecule in the traditional static sense but rather a dynamic
organelle. In 1970, Meyer et al. (21) were among the first
to suggest that glycogen was complexed with proteins and represented a
structural and functional unit of the muscle cells. Using TEM, they
estimated the diameter of this glycogen particle to be 20-30 m.
This value was in agreement with the average diameter (27 m) of
-glycogen particles previously described by Wanson and Drochmans
(31) in intact muscles. It is now known that the granule
is physically associated with a number of proteins, including glycogenin, glycogen synthase, glycogen phosphorylase, phosphorylase kinase, and phosphatase as well as scaffolding proteins [reviewed by
Roach et al. (24)]. This raises the distinct possibility that granules can be regulated regionally or even individually.
The term -particle has been applied to this protein-glycogen complex
and has been consistently described as a 20- to 30-m particle
(26). Several investigators have found the -particles to be located mainly in the subsarcolemmal and intermyofibrillar space,
in close proximity to the sarcoplasmic reticulum (SR) (9, 10, 26,
27, 29). Rybicka (26) reviewed the area and clearly
supported the concept that the structures commonly interpreted as
particles of glycogen actually represent dynamic organelles (glycosomes), which act as independent metabolic units with specialized functions. A heterogeneous distribution of glycosomes could result in
metabolic compartments with distinct characteristics and functions. For
example, biochemical and morphological evidence has demonstrated the
association of some glycosomes with the SR in both cardiac and skeletal
muscles (12, 21, 23, 31, 32). This subcellular location
may represent a metabolic compartment dedicated to SR function.
Ekblom and co-workers (9, 10, 29) have provided most of
our understanding of the compartmental distribution of skeletal muscle
glycogen in humans. They used TEM to describe qualitatively the
distribution of skeletal muscle glycogen in different subcellular locations, different muscle fiber types, and under different exercise conditions. They described glycogen as being stored in five
topographically different locations (10). They also
reported that exercise of different intensities metabolized glycogen
from specific subcellular locations and muscle fiber types and proposed
a selective pattern of subcellular glycogen metabolism, depending on
the exercise characteristics (9). Furthermore, they
described two separate and distinct populations of glycogen particles
with diameters of 22 and 56 m. They speculated that these
populations might represent the acid-soluble and -insoluble forms of
glycogen previously described by Jansson (16) [now
referred to as macro- and proglycogen, respectively (1, 2,
26)]. Although the results of these early studies are
interesting, the methodological approach used in estimating muscle
glycogen concentration lacked the objectivity and detailed information
of quantitative procedures. For example, their selection of fibers and
subcellular locations was not systematic, and the sample sizes were
restricted owing to the labor-intensive nature of TEM work and the lack
of advanced computerized processes to assist in data collection and analysis.
The purposes of the present study were to 1) develop and validate a TEM method to quantify the subcellular distribution of muscle glycogen to facilitate an objective evaluation of glycogen granules and 2) apply this TEM technique to resting human muscle biopsy samples to quantify the distribution of glycogen in different subcellular locations. We hypothesized that 1) the technique would provide data showing statistical differences in glycogen granule density in various subcellular locations and 2) the distribution of granule diameter would be bimodal, in accordance with pools of pro- and macroglycogen.
| |
METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Subjects
The study was approved by the University of Guelph's Human Ethics Committee and conformed to the standards set by the Declaration of Helsinki. Eleven male subjects were informed in writing about the nature of the study, volunteered to participate, and signed a consent form. Mean (±SE) age, height, weight, and maximal oxygen consumption (
O2max) of the
subjects were 25 ± 1 yr, 184 ± 2 cm, 86 ± 2 kg, and
53 ± 2 ml · kg
1 · min1, respectively.
Procedures
Resting muscle biopsies were obtained from the vastus lateralis of the subjects by use of the percutaneous needle biopsy technique (15). Each muscle sample was divided in two portions for biochemical and TEM analyses. Two independent investigators carried out the biochemical and TEM analyses.Biochemical Analysis
A piece of each muscle biopsy was immediately frozen in liquid N2 and subsequently stored at
80°C until it was freeze
dried and dissected free of visible blood, connective tissue, and other nonmuscle elements. A 1.5- to 3.5-mg portion of freeze-dried muscle was
extracted in 1.5 mM PCA and analyzed for pro- and macroglycogen (1). Enzymatic measurement of glucosyl units
(3) was then performed and reported as millimoles of
glucosyl units per kilogram of dry muscle weight.
TEM Analysis
The remaining fresh muscle tissue was immediately fixed in 2.0% glutaraldehyde in 0.1% sodium cacodylate buffer. The samples were then postfixed in a solution of 1% osmium tetroxide and 1.5% potassium ferricyanide in 0.1 M sodium cacodylate. The combination of osmium tetroxide with potassium ferricyanide was used for enhanced contrast-staining of intramuscular glycogen as previously described (14). The samples were then dehydrated in graded ethanol and subsequently infiltrated with graded mixtures of propylene oxide and Spurr's resin. Thin sections (~70m) were cut, and, for each sample, some sections were stained with uranyl acetate and lead citrate. After careful evaluation, it was concluded that the uranyl acetate and lead citrate staining were detrimental to glycogen analysis, and unstained sections were used for analysis. The
nonspecific nature of the chemical staining induced by uranyl acetate
and lead citrate renders essentially all cellular components
considerably electron dense and thus reduces the contrast between
glycogen particles and other cellular structures. The sections were
examined and photographed in a JEOL 1200 EX electron microscope. For
each session of electron microscopy, a standard calibration grid
(cross-grating replica from SPI Supplies, West Chester, PA; 2160 grid
lines/mm) was used to calculate the exact magnification.
A total of five to six muscle fibers were photographed from each muscle
sample (i.e., each subject). For each of the fibers per muscle biopsy,
a total of 11 images was obtained at ×20,000 magnification for
glycogen distribution analysis. Figure 1
illustrates the protocol used for the selection of the 11 systematic
sampling regions (images) within each fiber.
|
Our selection of different locations for the images (subsarcolemmal, superficial, and deep myofibrillar) was based on previous descriptions of the subcellular distribution of glycogen (9, 10, 27). For a given fiber (Fig. 1), images S1 to S5 comprised portions of the subsarcolemmal space: S1, one pole of a nucleus; S2, between the sarcolemma and a nucleus; S3, between a nucleus and the myofibrils; S4, sarcolemma adjacent to a cluster of mitochondria; and S5, sarcolemma lacking mitochondria. These sites were selected to gain representation of glycogen particles with respect to their association with the two main subsarcolemmal organelles, the nuclei and mitochondria. Images M1 to M6 were used to evaluate the myofibrillar space: M1, M3, and M5 represent myofibrils located immediately adjacent to the subsarcolemmal space ("superficial" myofibrillar space); M2, M4, and M6 were for analysis of myofibrils located toward the center of the muscle fiber diameter ("deep" myofibrillar space). Thus 55-66 images (×20,000 magnification) were digitized and analyzed for each of the 11 subjects/biopsies.
A preliminary analysis was performed to determine the minimal number of images necessary to represent the myofibrillar space of one whole fiber. This analysis examined the change in the coefficient of variation (standard deviation/mean × 100) obtained for the dependent variables (discussed in TEM analysis: dependent variables), as more images were included in the analysis (33). The coefficient of variation for the dependent variables decreased as the sampling increased from one to six images. There was no further decrease in the coefficient of variation with the addition of 7-10 data sets. On this basis, six myofibrillar samples were analyzed.
Digitizing
The images were digitized with the use of a cooled charge-coupled device camera and an Epi-illumination darkroom system (UVP, Upland, California). Each negative was digitized as two equal, half images to facilitate the analysis. A total of 22 images (11 negatives each divided into 2 images) was therefore used to represent each muscle fiber.Analysis of Subcellular Glycogen Distribution
The analysis was conducted in three parts: 1) single-particle evaluation, 2) total glycogen, and 3) intra- and intermyofibrillar glycogen with the use of image analysis software (Image Pro Plus, ver. 4.0). The calculations are described in detail in a later section, but the following is a brief overview of each part of the analysis.The single-particle analysis consisted of isolating the particles that were not clustered, determining the average single-particle diameter, and calculating the volume of a single particle. (There were no differences between the diameters of nonclustered and clustered particles.) These single-particle parameters were then compared between subcellular locations (subsarcolemmal space, superficial myofibrils, and deep myofibrils).
The total glycogen analysis consisted of determining both the total area of the image consisting of glycogen (glycogen area), as well as the number of glycogen particles that constituted this area (number of glycogen particles). The area was then converted to a volume (described below). Both the glycogen and the number of glycogen particles were compared between subcellular locations.
Finally, for both the deep and superficial myofibrillar images, the
intra- and intermyofibrillar glycogen analysis consisted of dividing
the amount of glycogen of each image into its respective intra- and
intermyofibrillar region (Fig. 2). The
sum of these data was the myofibrillar glycogen for that image.
|
Reliability Testing
Because the present technique is novel as a process to quantify the subcellular distribution of muscle glycogen, an extensive evaluation of its reliability and validity was essential. To test the reliability of the method, both intra- and interobserver reliability were evaluated. The former involved the main investigator performing analysis of two subjects (i.e., images of 10 different fibers from 2 different biopsies) twice, at least 2 wk apart. For the latter, an independent investigator performed the analysis of three subjects (i.e., images of 15 different fibers from 3 different biopsies), which were also analyzed by the main investigator.Our protocol to quantify the single-particle parameters did not differentiate between the intra- and intermyofibrillar particles. Therefore, before starting our single-particle analysis, it was necessary to ensure that the diameter of the single particles did not differ between these locations. Standard photographs and calibration grids were taken from three different samples [a subject with low muscle glycogen, a subject with high muscle glycogen, and a McArdle's disease (GSD V) patient with high muscle glycogen] at a magnification of ×40,000. An independent, blinded investigator manually outlined and traced individual glycogen particles in each of the intra- and intermyofibrillar spaces of different images from these samples. The MOP Videoplan program (Kontron, Germany) was precisely calibrated and used in this "manual" determination of the diameter of individual glycogen particles. A direct comparison was then performed between the intra- and intermyofibrillar particle diameters. Because the majority of the intermyofibrillar granules are arranged in clusters and the majority of the intramyofibrillar granules are isolated from each other (Fig. 2), this analysis also allowed elimination of the possibility of a difference between the size of particles arranged in clusters compared with those that were isolated.
Calculations
Biochemical analysis. Total glycogen concentration was calculated as the sum of pro- and macroglycogen (1). The percentage of total glycogen in the form of proglycogen (%PG) was determined by dividing the concentration of proglycogen by the total glycogen concentration and multiplying by 100 (%PG = proglycogen/total glycogen × 100).
TEM analysis: dependent variables.
For each muscle fiber analyzed with TEM, the following dependent
variables were obtained: 1) mean single-particle diameter (m) for deep myofibrillar space (Fig. 1: average of images M2, M4,
and M6), superficial myofibrillar space (Fig. 1: average of images M1,
M3, and M5), subsarcolemmal space (Fig. 1: average of images
S1-S5), and total fiber (images from all three spaces; Fig. 1:
average of images S1-S5 and M1-M6); 2) mean
single-particle volume (m3) for deep myofibrillar space,
superficial myofibrillar space, subsarcolemmal space, and total fiber;
3) number of glycogen particles (per µm3 of
muscle tissue) for deep myofibrillar space, superficial myofibrillar space, subsarcolemmal space, and total myofibrillar space (average of
the pooled deep and superficial myofibrillar space images; Fig. 1:
images M1-M6); and 4) glycogen volume
(µm3/µm3 of muscle tissue) for deep
myofibrillar space, superficial myofibrillar space, subsarcolemmal
space, and total myofibrillar space.
r3/3, where
is pi (3.141593) and r3 is (diameter/2)3.
Total glycogen area was divided by the mean single-particle area (both
values obtained by direct measurement from the digitized images) and
then by 0.070 µm (thickness of a thin section) to obtain the number
of glycogen particles per µm3 of muscle tissue. Glycogen
volume was calculated by multiplying the number of glycogen particles
(per µm3) by the mean single-particle volume
(µm3) to obtain the glycogen volume (in
µm3/µm3) of muscle tissue
(34).
The relative proportion of the myofibrillar glycogen located in the
intramyofibrillar region was calculated and compared between subcellular locations (deep and superficial myofibrils). A ratio of the
subsarcolemmal to total myofibrillar glycogen volume was calculated as
was the percent of total myofibrillar glycogen volume that was in the
intramyofibrillar space.
Figure 3 illustrates a summary of the
experimental design in which between 605 and 726 images were analyzed.
|
Intra- and interobserver reliability.
Intra- and interobserver reliability were evaluated by using these
calculations for each dependent variable: 1) the difference, 2) the %difference, 3) the absolute difference,
4) the %absolute difference, 5) the squared root
of the mean difference squared (RMS), 6) the %RMS,
7) the coefficient of correlation, and 8) t-test comparing the two respective groups of data. The
following is a summary of the formula used to calculate the above
mentioned criteria
|
![%Difference = [(<IT>X</IT><SUB>AVG</SUB> − <IT>Y</IT><SUB>AVG</SUB>) × 100]/[<IT>X</IT><SUB>AVG</SUB> + <IT>Y</IT><SUB>AVG</SUB>)/2]](/content/vol93/issue5/fulltext/1598/img002.gif)
|
|
|
![= [Absolute difference × 100]/[<IT>X</IT><SUB>AVG</SUB> + <IT>Y</IT><SUB>AVG</SUB>)/2]](/content/vol93/issue5/fulltext/1598/img005.gif)
|
![RMS = [&Sgr; (<IT>X</IT><SUB>i</SUB> − <IT>Y</IT><SUB>i</SUB>)<SUP>2</SUP>/<IT>n</IT>]<SUP>0.5</SUP>](/content/vol93/issue5/fulltext/1598/img006.gif)
|
![%RMS = [RMS × 100]/[<IT>X</IT><SUB>AVG</SUB> + <IT>Y</IT><SUB>AVG</SUB>)/2]](/content/vol93/issue5/fulltext/1598/img007.gif)
|
Statistics
Descriptive statistics were performed by using Microsoft Excel 1997. Inferential statistics were performed by using the SAS system (release 6.12). t-Tests were used to compare intra- and interobserver reliability. The multivariate repeated-measures design was assessed by using a mixed linear model with "subject" as the random effect and "subcellular locations and myofibrillar regions" as fixed effects. When these analyses revealed significant main effects or interactions, a Tukey's post hoc test was used to locate the pair-wise differences. Significance was considered to be P < 0.05. Data are presented as means ± SE.| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Biochemical Data
Subjects had an average total muscle glycogen concentration of 380 ± 41 mmol/kg dm, and 74 ± 2% was in the proglycogen form.Reliability Testing
Table 1 presents the results of the intra- and interobserver reliability testing for different variables. Our TEM technique of glycogen analysis is reliable and repeatable for a given examiner. The average percent difference between the two analyses was <2% for all the variables measured. In addition, the percentage of the RMS, a very conservative indicator of reliability, was <11% for each of the variables. Furthermore, a strong correlation, as indicated by r2, was significant for all the variables (P values not shown), and there were no significant differences between results of the two independent analyses performed by the main investigator.
|
Our TEM technique for glycogen analysis was reliable and repeatable between two independent investigators. The average percent difference between the results was <4% for all the variables measured.
Reliability was also evaluated by direct comparison between the intra-
and intermyofibrillar single-particle diameters that were determined
manually. The results for 3 different samples and more than 800 single
particles revealed no difference between the diameter of the intra- and
intermyofibrillar particles. In fact, the average diameter difference
between intra- and intermyofibrillar particles was found to be of <0.5
m (P = 0.25). Thus the data from the intra- and
intermyofibrillar space were combined during the single-particle
analyses. These results also suggest that the dimensions of granules
are not different if they are in clusters.
Dependent Variables
Figure 4 summarizes the results for single-particle analysis. There was a main effect of subcellular location (P < 0.0001). In addition, the average particle volume increased with depth in the fiber (Fig. 4). Post hoc analysis revealed that the subsarcolemmal particle volume (7,376 ± 48m3) was smaller than the superficial (8,462 ± 55 m3) and deep (8,802 ± 57 m3)
myofibrillar particle volumes (both P < 0.0001).
|
Figure 5 illustrates the distribution
frequency for the single-particle diameter for all fibers of the 11 subjects (n > 55,000 particles). Particle size appears
to be normally distributed and ranges from 10 to 44 m, with an
average size of ~25 m. Clearly there is no indication of two
distinct groups of granule sizes.
|
A main effect of subcellular location was present for both glycogen
volume (P < 0.0001) and number of glycogen particles
(P < 0.0001). Figure 6
illustrates the glycogen volume in each of the subcellular locations.
Although granule size was smaller in the subsarcolemmal space (Fig. 4),
the glycogen volume (µm3 of glycogen per
µm3 of muscle tissue) was 59 and 74% greater in the
subsarcolemmal space (0.037 ± 0.001) than in the deep
(0.023 ± 0.002) and superficial (0.021 ± 0.001) myofibrils,
respectively (both P < 0.001). Although glycogen
volume was 10% greater in the deep compared with the superficial
myofibrils, consistent with single-particle volume results (Fig. 4),
these differences did not reach statistical significance
(P = 0.32). The number of glycogen particles (as expressed in number per µm3 of muscle tissue) followed
the same pattern as the glycogen volume and was significantly greater
in the subsarcolemmal space (5,152 ± 332) than in the deep
(2,741 ± 146) and superficial (2,770 ± 190) myofibrils
(both P < 0.0001).
|
As expected, the intermyofibrillar region contained significantly more glycogen than the intramyofibrillar region. The former represented 87.8 ± 1.1% of the myofibrillar glycogen. This was observed for both the glycogen volume (P < 0.0001) and the number of glycogen particles (P < 0.0001), with the individual particle diameter not being different between regions (see above).
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
The main goals of the present study were to 1) develop and validate a TEM method for quantifying the subcellular distribution of muscle glycogen and 2) quantify the distribution of human skeletal muscle glycogen in different subcellular compartments. We hypothesized that the TEM technique would provide data that could be quantified to demonstrate objectively that glycogen granule distribution is heterogeneous within the cell and that granule size is bimodal. We have established a valid and reliable TEM technique to quantify the subcellular distribution of muscle glycogen. The major findings using this new method were 1) the glycogen distribution in human skeletal muscle is compartmentalized and heterogeneous, with the glycogen particles being more concentrated in the subsarcolemmal space than in the myofibrillar space; 2) single-particle diameter followed a continuous, normal distribution pattern; and 3) single-particle volume is greater in the myofibrillar space compared with that in the subsarcolemmal space.
Method Validity and Reliability
The TEM technique for quantifying intramuscular glycogen distribution is reliable both within and between investigators. The coefficients of variation for all variables were <2% for intraobserver reliability and <3% for interobserver reliability. This level of variation is less than the 5-10% that has been reported (1, 7) for repeated analysis using standard biochemical methods for measuring muscle glycogen. Essen and Henriksson (8) reported the coefficient of variation for biochemically determined glycogen values for two parts of the same fiber to be 9%. In addition, on examination of similar fibers from serial sections of the same muscle, White and Snow (33) found the coefficient of variation for a histochemical technique (periodic acid Schiff staining) to be 8%.Our method did not allow for a direct determination of total cellular glycogen as it was not possible to assess the volume relationship between the subsarcolemmal space, which has dense glycogen stores, and the total cell volume. It was not possible to measure individual muscle fiber circumference and volume, which is essential in calculating the relative contribution of the subsarcolemmal space to the total fiber volume. Because the diameter of individual muscle fibers is known to range from 10 to 100 µm (18), the error introduced by an average value would be tremendous. Eisenberg (6) reported that the subsarcolemnal space is ~1 µm wide. Thus, if one assumes that the fiber is round in cross section and that the diameter is 10, 50, or 100 µm, the proportion of the intracellular space that would be myofibrillar would be 64, 92, or 96%, respectively. Thus it is probable that myofibrillar glyogen represents the vast majority of carbohydrate stores.
The goal of this study was not to develop another method for measuring intramuscular glycogen. Our technique is extremely time consuming compared with traditional biochemical methods, but it provides the ability to quantitatively evaluate glycogen in subcellular locations, as well as the particle size and number.
In addition, a number of assumptions were necessary to convert glycogen
area into glycogen volume. These assumptions included that
1) the glycogen particles are spheres, 2) the
section thickness is 70 m, and 3) the sections for TEM have a
maximal thickness equal to the average particle diameter. Thus we
assumed that no particles are located on top of each other.
The first two assumptions are valid; glycogen particles have been
described consistently as regular spheroid organelles (12, 25,
26). The section thickness (70 m) was precise within ±10
m. With the section thickness of ~70 m and the average particle diameter of ~25 m, it is possible that, especially within the clusters, one to three glycogen particles could be located on top of
each other and escape detection. The two-dimensional TEM images could
not detect this juxtaposition. As detailed in METHODS, we
obtained glycogen volume by multiplying the number of particles in a
given image by the averaged single glycogen particle volume for that
image. By doing so, we limited the third dimension (thickness) to the
diameter one particle. We initially tried to evaluate the optical
density of the glycogen particles. We compared the density of clusters
to that of a single glycogen particle by assuming that clusters would
be denser due to the effect of overlapping particles. However, a single
isolated glycogen particle already displayed the maximal density. Given
this limitation, we elected for the conservative protocol of
restricting the third dimension (thickness) to the maximal particle
diameter for each particular image.
The present method is the first objective, quantitative technique that has been applied to glycogen distribution in human skeletal muscle. We have demonstrated that the technique is valid and reliable. However, it could be improved in several ways. As discussed above, granule overlapping is a limitation. In addition, when one is examining a three-dimensional structure with varying diameters, there is not a uniform probability that all particles will be sectioned and observed (13). To address this, two consecutive serial samples could have been taken. Finally, the fibers should have been fixed at resting length because shortening may have resulted in some spacial disturbance and overlap of granules.
Subcellular Locations
We found muscle glycogen to be heterogeneously distributed within the cell in a manner similar to what has been described qualitatively previously (9, 10, 27, 29). Schmalbruch and Kamieniecka (27) reported glycogen granules to be either closely packed in strands marking the boundaries of the myofibrils in the I-band region or to be located in rows or as single granules within the myofibrils (see Fig. 2). Sjostrom et al. (29) extended these observations by reporting a third subcellular site of glycogen storage, the subsarcolemmal space. They characterized the intramuscular distribution of glycogen of their control subjects (i.e., resting) as 1) large subsarcolemmal accumulations, 2) large accumulations between the myofibrils (especially at I-band in close proximity to the mitochondria and SR), and 3) relatively smaller intramyofibrillar accumulations. They also reported that the intramyofibrillar glycogen particles were oriented as if they occupied a space previously filled with a thin or thick myofilament. In addition, Friden and colleagues (10) further emphasized an I-band-based location of glycogen, both for intra- and intermyofibrillar glycogen, and noted that the intramyofibrillar glycogen granules were found both in horizontal rows beginning at the lateral border of the I band and extending toward the M line as well as longitudinal rows of single particles bordering the Z line. These observations were apparent in the present study (see Fig. 2).These earlier experiments were insightful, but the methodological approach used in estimating muscle glycogen concentration was strictly qualitative. Because of the labor-intensive nature of TEM methods and the lack of advanced computerized processes to assist in data collection, these experiments' sample sizes were small, and sampling was not performed systematically. For example, in the investigations of human skeletal muscle (9, 10, 29), the resting controls were four or five subjects, and in one case (10) there were only three subjects who exercised. Furthermore, the number of fibers sampled was rarely stated nor was there a detailed description of how subcellular sampling sites were determined.
In the present study, we provide the first quantitative analysis of subcellular glycogen location. We confirmed the presence of and quantified the characteristics of three morphologically distinct compartments in human skeletal muscle: subsarcolemmal, intramyofibrillar, and intermyofibrillar spaces. We found that the intermyofibrillar space is the largest glycogen compartment, although the subsarcolemmal space is more densely packed with glycogen granules, having almost twice the granule number per unit area than the myofibrillar space (Fig. 6). Our results also confirm that the particles located between the myofilaments (intramyofibrillar region) were either isolated from each other or aligned in strands but never in clusters.
These anatomical compartments probably have metabolic significance. Friden and colleagues (9) compared the pattern of glycogen depletion during a marathon to that during an intense, anaerobic, sprint exercise. They found the subsarcolemmal glycogen fraction was not depleted at the end of a marathon, whereas the other locations were "empty." In contrast, in fibers of the three (10) and six (9) subjects who performed sprints, there was a marked depletion of subsarcolemmal pools, whereas the "perimitochondrial" locations remained intact at fatigue. They concluded that, depending on the type of exercise, differential sequential glycogen utilization patterns can be observed, and suggested a compartmentalized metabolism of glycogen. The present technique will facilitate evaluation of their hypothesis that the three subcellular glycogen storage sites represent distinct metabolic compartments.
Our findings represent the first evidence of a graded increase in single glycogen particle size from the sarcolemma to the center of the muscle fiber (Fig. 4). The metabolic implications of this finding are presently unknown. This may be somewhat analogous to the previous report of a gradient of intermyofibrillar mitochondrial content, from high to low, from the sarcolemma to the center of the muscle fiber (17). The intermyofibrillar and subsarcolemmal mitochondrial fractions have been shown to have distinct biochemical properties and differential adaptations to increasing or decreasing muscle activity (4).
Regression analyses (data not shown) between biochemical and TEM data also support the concept of metabolic compartmentalization. As noted earlier, the major portion of glycogen is myofibrillar. The biochemically determined total glycogen had a strong correlation with the myofibrillar value (r2 = 0.60). Proglycogen (400,000 Da), determined biochemically, had a stronger correlation with the subsarcolemmal glycogen (r2 =0.61) than with the myofibrillar one (r2 =0.48). In contrast, macroglycogen (400,000 to 107 Da) correlated better with the myofibrillar (r2 = 0.55) than with the subsarcolemmal (r2 = 0.34) glycogen fraction. This is consistent with our finding that single-particle size in the subsarcolemmal space was significantly smaller than that in the myofibrillar space. Further studies are required to evaluate these suggestions.
Single-Particle Distribution
The present study demonstrates for the first time that, in resting human skeletal muscle, glycogen particles present as a continuum of sizes ranging from 10 to 44m in diameter (Fig. 5). This is in
contrast to our hypothesis, which was based on the description of
Friden et al. (10), of two distinct populations of
glycogen particles with average diameters of 22 and 56 m. However,
Friden et al.'s results are based on the analysis of 144 glycogen
particles, whereas we measured >55,000 granules. The mean single
glycogen particle diameter was found to be 25 m, which was in
agreement with values previously reported (9, 12, 25-27,
31). Not only did we not find evidence for a bimodal distribution, we did not observe any particles larger than 44 nm. The
frequency distribution of the single-particle diameters from 11 subjects suggests that the distribution is normal, although slightly
skewed on the left. The latter finding is to be expected, considering
the error introduced by the 70-m sections relative to the 25-m
average diameter of the glycogen particles. If >50% of a spherical
granule is not included in the section, then the true diameter will be
underestimated. Conversely, if >50% of such a particle is included,
the correct diameter will be observed. This bias can only lead to an
underestimation of a particle diameter, hence the distribution being
skewed on the left. As reported by Williams (34), this
only becomes a significant problem when the object being measured has a
diameter greater than the thickness of the thin section.
We found that the maximal granule diameter was 44 m. The extent to
which this finding agrees with the values predicted from mathematical
modeling is striking. Considerable evidence converges to propose that
an optimally efficient glycogen -particle would have a maximal
diameter of 42 m (11, 19, 20). According to the models,
the "mature" glycogen -particle (macroglycogen) would be a
spherical particle arranged in 12 concentric layers (tiers) of
carbohydrate, with a molecular mass of 107 Da. Each tier is
estimated to add 3.8 m to the diameter (11), and
the amount of glucose available in any outer tier is always 34.6% of
the total amount of glucose in the particle (20).
This means that what appear to be small differences in granule size, especially in the larger granules, could be quantitatively very important with regard to carbohydrate storage.
There is controversy in the literature as to whether proglycogen and
macroglycogen only exist as discrete entities (2) or as a
wide range of molecular weights (30). Our data firmly support the theory that there is a continuous spectrum of glycogen particles sizes, ranging from sizes considerably smaller than the upper
limit of proglycogen (~30 m) to sizes corresponding to the maximal
macroglycogen diameter (~42 m).
Fiber Types
Although fiber-type comparisons were not a main goal of the study, a preliminary examination was conducted. It has been demonstrated that mammalian Z lines vary in thickness with muscle fiber type (10, 18, 22, 27, 29). Although the values reported vary slightly between studies, it is clear that, in any particular muscle, the widest Z lines are found in the type I fibers and the narrowest in the type IIB fibers [reviewed by Landon (18)]. Eisenberg (6) concluded that in any one preparation, the upper and lower limits of the measured range of Z-line values may be expected to correspond to type I and type IIB fibers, respectively, but that other independent criteria are required to identify fibers containing Z lines of intermediate widths. On the basis of this, we used Z-line width as our primary criterion, and in cases where there was uncertainty, three independent investigators determined the fiber type based on the following criteria from Landon (18): 1) lipid content, 2) shape and content of the mitochondria, and 3) SR and T-tubules appearance. With the use of this technique, mean ± SE of Z-line width for type I, IIA, and IIB were 92.8 ± 2.8, 79.8 ± 1.2, and 70.0 ± 0.6 nm, respectively. These values are in agreement with those reported previously (10, 18, 22, 27) and suggest that indeed the classifications were correct. However, given the limited number of fibers that were processed, our findings regarding fiber-type differences should be viewed as preliminary. Furthermore, very few type IIB fibers were identified, and thus we only statistically compared type I and IIA fibers. This aspect needs further investigation, which should include more accurate identification of fiber types and the inclusion of more fibers.The mean particle volume for type I fibers was significantly less than that for type IIA fibers (7,018 ± 565 vs. 8,466 ± 517 nm3, respectively). One could speculate that because type I fibers are being constantly recruited during activities of daily living, their glycogen particles are in a more dynamic state of turnover and thus are smaller. This difference may seem modest, but the difference in diameter (24.5 ± 0.9 vs 25.3 ± 0.5 nm, respectively) is almost 33% of a tier of glycogen, and the addition of a tier represents a 50% increase in carbohydrate.
Subsarcolemmal-to-total myofibrillar glycogen ratio, although not significantly different between fiber types, was 21% greater in type IIA compared with type I fibers. This suggests that there may be relatively less subsarcolemmal glycogen in type I than in type IIA fibers.
Friden and co-workers (10) suggested that the only consistent difference between type I and II fibers was that more glycogen particles were located at the H zone in the type II fibers. We did not differentiate between the different intramyofibrillar locations. However, we found that type I fibers had a significantly greater proportion of myofibrillar glycogen located in the intramyofibrillar space (14.3 ± 0.9 vs. 9.7 ± 0.7%, respectively). This finding is in agreement with that of Schmalbruch and Kamieniecka (27), who observed that a larger proportion of the glycogen in type I fibers was located between the actin and myosin filaments, whereas the glycogen in type IIA and IIB fibers was preferentially distributed between myofibrils. Thus we are able to substantiate these earlier descriptions with quantified results.
In conclusion, this study presents a novel method to quantify the
subcellular distribution of skeletal muscle glycogen. The method was
found to be valid and reliable, both within and between investigators.
We found that, in resting, human, skeletal muscle, 1) the
glycogen distribution is compartmentalized and heterogeneous, with the
glycogen being more concentrated in the subsarcolemmal space than in
the myofibrillar space, and the single-particle volume was greater in
the latter; and 2) single-particle sizes range from 10 to 44 m (25.2 ± 2.8 m) and follow a normal distribution. The
method also has the potential to be used to study fiber-type differences. A preliminary examination suggested that there were differences in the subcellular distribution of glycogen between fiber
types, the single particles being bigger in type II fibers and the
proportion of intramyofibrillar glycogen being greater in type I. The
results of the present study represent the first quantitative
demonstration of a compartmentalized pattern of subcellular glycogen
deposition in human skeletal muscles.
| |
ACKNOWLEDGEMENTS |
|---|
We express appreciation to Drs. Jacqueline Bourgeois and George Harauz for expertise in electron microscopy and critical input into the development of the methodology. The technologists from the electron microscopy department of McMaster University deserve special thanks for time and patience during the method development.
| |
FOOTNOTES |
|---|
This study was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Gatorade Sports Science Institute. I. Marchand held a NSERC scholarship, and J. Shearer held an industrial NSERC scholarship in association with the Gatorade Sports Science Institute.
Address for reprint requests and other correspondence: T. E. Graham, Dept. of Human Biology and Nutritional Sciences, Univ. of Guelph, Guelph, Ontario, Canada N1G 2W1 (E-mail: terrygra{at}uoguelph.ca).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
July 12, 2002;10.1152/japplphysiol.00585.2001
Received 6 June 2001; accepted in final form 11 July 2002.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
1.
Adamo, KA,
Tarnopolsky MA,
and
Graham TE.
Dietary carbohydrate and postexercise synthesis of proglycogen and macroglycogen in human skeletal muscle.
Am J Physiol Endocrinol Metab
275:
E229-E234,
1998
2.
Alonso, MD,
Lomako J,
Lomako WM,
and
Whelan WJ.
A new look at the biogenesis of glycogen.
FASEB J
9:
1126-1137,
1995[Abstract].
3.
Bergmeyer, HU,
Bernt E,
Schmidt F,
and
Stork H.
D-Glucose determination with hexokinase and glucose-6-phosphate dehydrogenase.
In: Methods of Enzymatic Analysis, edited by Bergmeyer HU.. New York: Academic, 1974, p. 1196-1201.
4.
Cogswell, AM,
Stevens RJ,
and
Hood DA.
Properties of skeletal muscle mitochondria isolated from subsarcolemmal and intermyofibrillar regions.
Am J Physiol Cell Physiol
264:
C383-C389,
1993
5.
Drochmans, P.
Etude au microscope electronique de colorations negatives du glycogene particulaire. Morphologie du glycogene.
J Ultrastruct Res
6:
141-163,
2001.
6.
Eisenberg, BR.
Quantitative ultrastructure of mammalian skeletal muscle.
In: Handbook of Physiology. Skeletal Muscle Bethesda, MD: Am. Physiol. Soc, 1983, sect. 10, chapt. 3, p. 73.
7.
Esbjornsson-Liljedahl, M,
Sundberg CJ,
Norman B,
and
Jansson E.
Metabolic response in type I and type II muscle fibers during a 30-s cycle sprint in men and women.
J Appl Physiol
87:
1326-1332,
1999
8.
Essen, B,
and
Henriksson J.
Glycogen content of individual muscle fibres in man.
Acta Physiol Scand
90:
645-647,
1974[ISI][Medline].
9.
Friden, J,
Seger J,
and
Ekblom B.
Implementation of periodic acid-thiosemicarbazide-silver proteinate staining for ultrastructural assessment of muscle glycogen utilization during exercise.
Cell Tissue Res
242:
229-232,
1985[ISI][Medline].
10.
Friden, J,
Seger J,
and
Ekblom B.
Topographical localization of muscle glycogen: an ultrahistochemical study in the human vastus lateralis.
Acta Physiol Scand
135:
381-391,
1989[ISI][Medline].
11.
Goldsmith, E,
Sprang S,
and
Fletterick JR.
Structure of Maltoheptaose by difference Fourier methods and a model for glycogen.
J Mol Biol
156:
411-427,
1982[ISI][Medline].
12.
Goldstein, MA,
Murphy DL,
Van Winkle WB,
and
Entman ML.
Cytochemical studies of a glycogen-sarcoplasmic reticulum complex.
J Muscle Res Cell Motil
6:
177-187,
1985[ISI][Medline].
13.
Gundersen, HJG,
Bagger P,
Bendtsen TF,
Evans SE,
Korbo L,
Marcussen N,
Moller A,
Nielsen K,
Nyengaard JR,
Pakkerberg B,
Sorensen FB,
Vesterby A,
and
West MJ.
The new stereological tools: disector, fractionator, nucleator and point sampled intercepts and their use in pathological research and diagnosis.
APMIS
96:
857-881,
1988[ISI][Medline].
14.
Hayat, MA.
Principles and Techniques of Electron Microscopy. Boca Raton, FL: CRC, 1989.
15.
Hultman, E.
Muscle glycogen in man determined in needle biopsy specimens.
Scand J Clin Lab Invest
113:
337-340,
1967.
16.
Jansson, E.
Acid soluble and insoluble glycogen in human skeletal muscle.
Acta Physiol Scand
113:
337-340,
1981[ISI][Medline].
17.
Kayar, SR,
Hoppeler H,
Claassen H,
and
Oberholzer F.
Acute effects of endurance exercise on mitochondrial distribution and skeletal muscle morphology.
Eur J Appl Physiol
54:
578-584,
1986.
18.
Landon, DN.
Skeletal muscle: normal morphology, development and innervation.
In: Skeletal Muscle Pathology, edited by Mastaglia FL,
and Detchant LW.. London: Churchill Livingstone, 1992, p. 1-40.
19.
Melendez, R,
Melendez-Hevia E,
and
Cascante M.
How did glycogen structure evolve to satisfy the requirement for rapid mobilization of glucose? A problem of physical constraints in structure building.
J Mol Evol
45:
446-455,
1997[ISI][Medline].
20.
Melendez-Hevia, E,
Waddell TG,
and
Shelton ED.
Optimization of molecular design in evolution of metabolism: the glycogen molecule.
Biochem J
295:
477-483,
1993.
21.
Meyer, F,
Heilmeyer LMG,
Haschke RH,
and
Fisher EH.
Control of phosphorylase activity in a muscle glycogen particle: isolation and characterization of the protein-glycogen complex.
J Biol Chem
245:
6642-6648,
1970
22.
Payne, CM,
Stern LZ,
Curless RG,
and
Hannapel LK.
Ultrastructural fibre typing in normal and diseased human muscle.
J Neurol Sci
25:
99-108,
1975[ISI][Medline].
23.
Polishchuk, SV,
Brandt NR,
Meyer H,
Varsanyi M,
and
Heilmeyer LMG, Jr.
Does phosphorylase kinase control glycogen biosynthesis in skeletal muscle?
FEBS Lett
362:
271-275,
1995[ISI][Medline].
24.
Roach, PJ,
Cheng C,
Huang D,
Lin A,
Mu J,
Skurat AV,
Wilson W,
and
Zhai L.
Novel aspects of the regulation of glycogen storage.
J Basic Clin Physiol Pharmacol
9:
139-151,
1998[Medline].
25.
Rybicka, KK.
Simultaneous demonstration of glycogen and protein in glycosomes of cardiac tissue.
J Histochem Cytochem
29:
4-8,
1981[Abstract].
26.
Rybicka, KK.
Glycosomes: the organelles of glycogen metabolism.
Tissue Cell
28:
253-265,
1996[ISI][Medline].
27.
Schmalbruch, H,
and
Kamieniecka Z.
Fibre types in the human brachial biceps muscle.
Exp Neurol
44:
313-328,
1974[ISI][Medline].
28.
Scott, RB,
and
Still WJS
Glycogen in human peripheral blood leucocytes. II. The macromolecular state of leucocyte glycogen.
J Clin Invest
47:
353-359,
1968[ISI][Medline].
29.
Sjostrom, M,
Friden J,
and
Ekblom B.
Fine structural details of human muscle fibres after fibre type specific glycogen depletion.
Histochemistry
76:
425-438,
1982[ISI][Medline].
30.
Skurat, AV,
Lim SS,
and
Roach PJ.
Glycogen biogenesis in rat 1 fibroblasts expressing rabbit muscle glycogenin.
Eur J Biochem
245:
147-155,
1997[ISI][Medline].
31.
Wanson, JC,
and
Drochmans P.
Rabbit skeletal muscle glycogen. A morphological and biochemical study of glycogen b-particles isolated by precipitation-centrifugation method.
J Cell Biol
38:
130-150,
1968
32.
Wanson, JC,
and
Drochmans P.
Role of the sarcoplasmic reticulum in glycogen metabolism. Binding of phosohorylase, phosphorylase kinase, and primer complexes to the sarcovesicles of rabbit skeletal muscle.
J Cell Biol
54:
206-224,
1972
33.
White, MG,
and
Snow DH.
Quantitative histochemical study of glycogen depletion in the maximally exercised Thoroughbred.
Equine Vet J
19:
67-69,
1987[ISI][Medline].
34.
Williams, MA.
Quantitative methods in biology.
In: Practical Methods in Electron Microscopy. Cambridge, UK: North-Holland, 1977, vol. 6, p. 68.
This article has been cited by other articles:
|
|
 |
|
  T. A. Duhamel, H. J. Green, R. D. Stewart, K. P. Foley, I. C. Smith, and J. Ouyang Muscle metabolic, SR Ca2+-cycling responses to prolonged cycling, with and without glucose supplementation J Appl Physiol, December 1, 2007; 103(6): 1986 - 1998. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Wilson, J. E. Gusba, D. L. Robinson, and T. E. Graham Glycogenin protein and mRNA expression in response to changing glycogen concentration in exercise and recovery Am J Physiol Endocrinol Metab, June 1, 2007; 292(6): E1815 - E1822. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Marchand, M. Tarnopolsky, K. B. Adamo, J. M. Bourgeois, K. Chorneyko, and T. E. Graham Quantitative assessment of human muscle glycogen granules size and number in subcellular locations during recovery from prolonged exercise J. Physiol., April 15, 2007; 580(2): 617 - 628. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. A. Duhamel, J. G. Perco, and H. J. Green Manipulation of dietary carbohydrates after prolonged effort modifies muscle sarcoplasmic reticulum responses in exercising males Am J Physiol Regulatory Integrative Comp Physiol, October 1, 2006; 291(4): R1100 - R1110. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Shearer, R. Wilson, and T. Graham Rebuttal to Abram Katz's Letter To The Editor Am J Physiol Endocrinol Metab, April 1, 2006; 290(4): E758 - E759. [Full Text] [PDF]
|
|
|
|
|
|
J. Shearer, R. J. Wilson, D. S. Battram, E. A. Richter, D. L. Robinson, M. Bakovic, and T. E. Graham Increases in glycogenin and glycogenin mRNA accompany glycogen resynthesis in human skeletal muscle Am J Physiol Endocrinol Metab, September 1, 2005; 289(3): E508 - E514. [Abstract] [Full Text] [PDF]
|
|
| ||||
