Abstract
Quantification of a specific muscle mRNA per total RNA (e.g., by Northern blot analysis) plays a crucial role in assessment of developmental, experimental, or pathological changes in gene expression. However, total RNA content per gram of a particular fiber type may differ as well. We have tested this possibility in the distinct fiber types of adult rat skeletal muscle. Sections of single fibers were hybridized against 28S rRNA as a marker for RNA content. Quantification of the hybridization showed that the 28S rRNA content decreases in the order i>IIA>IIX>IIB, where Type I fibers show a five- to sixfold higher expression level compared to Type MB fibers. Results were verified with an independent biochemical determination of total RNA content performed on pools of histochemically defined freeze-dried single fibers. In addition, the proportion of myosin heavy chain (MHC) mRNA per μg of total RNA was similar in slow and fast fibers, as demonstrated by Northern blot analysis. Consequently, Type I fibers contain five- to sixfold more MHC mRNA per μg of tissue than IIB fibers. These differences are not reflected in the total fiber protein content. This study implies that proper assessment of mRNA levels in skeletal muscle requires evaluation of total RNA levels according to fiber type composition. (J Histochem Cytochem 47:995–1004, 1999)
I
The several distinct MHC isoforms, which can even be co-expressed in the same fiber, provide the muscle fibers with the ability to modulate their contractile characteristics by switching different isoform genes on and off. This property, commonly referred to as muscle plasticity, has been extensively studied in experimental and pathological conditions, mainly at the protein level (Swyngedauw 1980; Pette and Staron 1990; Caiozzo et al. 1992, 1996). The introduction of molecular biological techniques, in addition to biochemical techniques, enables the detection of changes in isoform expression at the transcriptional level. The vast majority of studies that assessed changes in gene expression have quantified muscle mRNA in tissue homogenates using Northern blot, dot-blot, RNase protection, or RT-PCR techniques. As a consequence, cellular resolution is lost and the relative changes in specific mRNAs were expressed in relation to the amount of total RNA (specific mRNA/total RNA). However, no account was taken of whether the total amount of RNA per amount of tissue varied among the distinct fiber types or whether it varied under different experimental and pathological conditions. Clearly, without taking into account such differences in total RNA content, absolute increases in signal can be misleading with respect to the effect of experimental interventions, e.g., exercise, chronic stimulation, on different populations of muscle fiber types. Therefore, in this study we examined the total RNA content using in situ hybridization (ISH) and densitometry on histochemically defined single fibers (Sant'Ana Pereira et al. 1995b,b; Jonker et al. 1997). The densitometric results were compared to a biochemical determination of total RNA content in the same fiber. ISH on whole-muscle cross-sections and Northern blot analysis were carried out to illustrate the results from the single-fiber experiments.
Our data show that Type I fibers have a five- to sixfold higher total RNA content than the fastest Type IIB fibers, whereas the MHC mRNA content per μg of total RNA is similar. Consequently, slow fibers also have five- to sixfold more MHC mRNA per μg of tissue than Type IIB fibers.
Material and Methods
Animal Care
Wistar rats (250-300 g) were housed with a 12-hr light and 12-hr dark cycle, with food and water available ad libitum. After decapitation, the medial gastrocnemius (MG), soleus, and extensor digitorum longus (EDL) muscles were carefully excised. The muscles were slightly stretched, stored for 15 min at 4C, frozen in liquid nitrogen-cooled isopentane, and prepared for cutting of whole-muscle cross-sections as described (Sant'Ana Pereira et al. 1995b). To dissect single fibers, frozen whole-muscle preparations were freeze-dried (−60C) and processed under standard conditions of temperature and humidity as described by Sant'Ana Pereira et al. (1995a). Each fiber was cut into two parts. The first part was embedded in RNase-free gelatin solution and used to cut transverse serial sections (10 μm), which were processed for histochemistry and ISH. The second part was used for total RNA determination using the orcinol method (see below).
Histochemistry
Serial cryostat sections of whole-muscle and gelatin-embedded fibers were stained for acid- and alkaline-labile myofibrillar adenosine triphosphatase (mATPase) activity and for an additional mATPase assay that solely differentiates fibers expressing MHC IIX (Sant'Ana Pereira et al. 1995b, 1997). The combination of these staining procedures allows correct identification of pure and hybrid I, IIA, IIX and IIB fibers. Further analyses of single fibers were restricted to pure fibers to rule out, as far as possible, differences caused by the co-expression of two or more distinct MHC isoforms.
Preparation of cRNA Probes
Single [35S]-CTP and double [35S]-CTP and [35S]-UTP-labeled sense and antisense cRNAs of rat β (Boheler et al. 1992), IIA (De Nardi et al. 1993), IIX (De Nardi et al. 1993), and IIB (Schiaffino et al. 1989a) MHC isoforms, mouse 28S rRNA, human β-MHC ATP binding site (Jaenicke et al. 1990) and the mouse elongation factor-1α (EF-1α) were generated according to standard protocols (Melton et al. 1984). The EF-1α probe was isolated from an embryonic mouse cardiac cDNA library and corresponds to the bases 230-1319 of EF-1α mRNA (accession number X13661). To compensate for the differences in length between the fast MHC isoform probes and the slow β-MHC probe (414 nt for the β-MHC probe, and 120, 121, and 76 nt for the IIA, IIX, and IIB MHC probes, respectively), the fast MHC isoform specific probes were double labeled and the exposure time was twice as long compared to the single labeled β-MHC probe, which was exposed for 1 week. The 28S rRNA probe was a single labeled cRNA, transcribed from the 930-bp fragment of the mouse 28S rRNA cDNA corresponding to bases 3660–4590 of 28S rRNA (accession number X00525). The human β-MHC ATP binding site probe was a double labeled cRNA transcribed from the 183-bp fragment of the human β-MHC cDNA corresponding to nucleotides 460-643 of β-MHC mRNA (accession number M25137). The mouse EF-1α probe was a double labeled cRNA transcribed from the 1089-bp fragment of the mouse EF-1α cDNA.
In Situ Hybridization
Cryostat sections of whole-muscle preparations and gelatin-embedded single fibers were collected on RNase-free aminopropyltriethoxysilane (AAS)-coated slides and kept at −80C. After freeze-drying (−60C) for 2-3 hr, the sections were brought to room temperature (RT), fixed for 20 min in 4% paraformaldehyde (PFA) in PBS, rinsed in PBS, dehydrated through graded ethanol steps, and dried in a filtered air stream.
ISH was performed essentially according to Moorman et al. (1993). Briefly, sections were pretreated as follows: 20 min 0.2 N HCl, 5 min bidistilled water, 2 × SSC (SSC is 0.15 M NaCl, 0.015 M sodium citrate, pH 7.2) for 10 min at 70C, 5 min in bidistilled water, 20 min digestion at 37C in 0.1% pepsin dissolved in 0.01 N HCl, 30 sec in 0.2% glycine in PBS, twice for 30 sec in PBS, 20 min of postfixation in a 4% freshly made formaldehyde solution for 20 min, 5 min in bidistilled water, 5 min in 10 mM EDTA, 5 min in 10 mM dithiotreitol (DTT), and finally the sections were dried in a filtered air stream. The hybridization mixture contained 50% formamide, 10% dextran sulfate, 2 × SSC, 2 × Denhardt's solution, 0.1% Triton X-100, 10 mM DTT, 200 ng/μl heat-denatured herring sperm DNA. Hybridization was performed overnight at 52C and the sections were washed as follows: a rinse in 1 × SSC, 30 min in 50% formamide/1 × SSC at 52C, 10 min 1 × SSC at RT, 30 min at 37C in RNase A (10 μg/ml), 10 min in 1 × SSC, and 10 min in 0.1 × SSC. The sections were dried, dipped in an autoradiographic emulsion (Ilford Nuclear Research Emulsion G-5), and exposed for 2 days (28S rRNA), 1 week (β-MHC and MHC ATP binding site mRNA), or 2 weeks (IIA, IIX, IIB MHC mRNA and EF-1α mRNA). Development time was set at 4 min. Controls were performed on serial sections under identical conditions either in the presence of the sense probe or after pretreatment of the sections with RNase.
Biochemical Determination of Total RNA
In addition to the ISH with the 28S rRNA probe, a biochemical method was used to determine the total RNA content. Fragments of histochemically characterized single fibers were pooled according to their MHC composition (i.e., I, IIA, IIX, and IIB) and the total amount of tissue (200-300 μg) was determined using a 0.1 μg rated Sartorius-(S4) supramicrobalance. A scaled down version of the orcinol assay (Munro 1966) was used essentially as described by van den Hoff et al. (1997). Briefly, the samples were homogenized in ice-cold sterile bidistilled water. The nucleic acids were precipitated in 0.5 M HClO4 and hydrolyzed by incubation in 0.3 M KOH at 60C for 1 hr. After removal of DNA by 0.5 M HClO4 precipitation, the samples were boiled for 30 min in 6 M HCl, 0.01% FeCl3, and 0.3% orcinol.
Optical density (OD) was measured at a wavelength of 660 nm and RNA concentrations were calculated using a calibration curve with total adult liver RNA isolated by cesium chloride ultracentrifugation (Chirgwin et al. 1979).
Image Recording
Images of the fiber sections were recorded as described in detail by Jonker et al. (1997). Briefly, a Photometrics cooled CCD camera (Tucson, AZ) attached to an Axioplan microscope (Zeiss; Oberkochen, Germany) equipped with a stabilized power supply and a monochromatic filter was used to record the images. The resolution of the images was set at 512 × 512 pixels, which corresponds to 175 × 175 μm2 in the object plane at X40 magnification. The acquired images were converted to OD images using the formula
where I is the transmitted light (the acquired image) and IO the incident light (the image of the light source).
Image Analysis
Images were analyzed with the software package NIH-Image (v1.61, Rasband, National Institute of Health, Bethesda, MD; a public domain image processing and analysis program). First, a mask of the cell border was obtained by manual tracing of the fiber contour in a serial section used for mATPase staining (Figure 1A). As a second step, this contour was manually positioned on the corresponding ISH section of the fiber (Figure 1B). Subsequently, a distance transformation was performed from the positioned contour inward to the core of the fiber and outward towards the periphery of the image. This resulted in two distance-transformed images combined in Figure 1C.
From both distance-transformed images, consecutive zones of five gray values width (representing steps of 1.65 μm) were selected. For reasons of clarity, only two zones are highlighted in Figure 1D. Subsequently, the OD image of the ISH section was masked with each zone (Figure 1E; as indicated for Figure 1D, masking is shown for only two zones). In this masked selection, the mean OD and the area of each zone were measured, resulting in a graphic representation of the density distribution of each fiber (Figure 1F).
The calculation of the progressive mean OD, starting at the outermost zone, was used to determine the background density level. When the progressive mean OD increased by more than 5% in one step, the previous calculated mean OD was considered to be the background (indicated by the gray line in Figure 1F). Measured mean OD values per zone were corrected by subtraction of the background and multiplication by the area of the zone. The resulting OD values were summed to calculate the integrated OD (IOD) per fiber. IOD values for each fiber were then divided by cross-sectional area to correct for differences in fiber size. The corrected IOD is related to the 28S rRNA concentration in the fiber, because IOD is linearly related to the amount of radioactivity bound to the section due to specific hybridization, which in turn is related to the amount of RNA in the section (Jonker et al. 1997).
Statistical Analysis
The data are presented as mean ± SEM. A Kruskal-Wallis nonparametric one-way analysis of variance (ANOVA) followed by a multiple comparison of groups was used to test for differences among the distinct fiber types (Conover 1980). The significance level was set at p<0.05.
Northern Blot Analysis
Total RNA was isolated from soleus (more than 90% Type I fibers) and EDL (more than 90% Type II fibers) muscles according to Chomczynski and Sacchi (1987). Ten μg of each total RNA preparation was electrophoresed, blotted onto a nitrocellulose membrane (Hybond; Amersham, Poole, UK), and hybridized using a 32P-labeled random-primed probe of the 183-nt ATP binding site corresponding to nucleotides 460-643 of the human β-MHC cDNA (accession number M25137) (Sambrook et al. 1989). After overnight hybridization and washing of the membranes, blots were exposed in a Phosphor Image analyzer (Molecular Dynamics; Sunnyvale, CA) and analyzed using the Image Quant 3.3 program (Molecular Dynamics). Ribosomal RNA (18S and 28S) staining was used as the internal standard for the amount of RNA loaded on the gel.
Analysis and quantification of ISH on serial cryostat sections of gelatin-embedded rat single-muscle fibers. A section used for mATPase staining was used to trace the fiber contour (A). This contour was positioned manually on the corresponding ISH sections (B). Taking the cell border as a starting point, distance transformations were performed towards the core of the fiber and towards the periphery of the image. During this procedure, the distance from the cell border is converted into consecutive gray values (C). Subsequently, successive zones of five gray values width were selected, two of which are highlighted (D). Each zone is masked with the OD image of the ISH section (E; masking is shown only for the same two zones presented in D). Mean OD and area of each zone are calculated and can be represented as a density distribution of the fiber (F), in which the white line represents the border of the cell. The gray and white bars represent the outer zones in D and E. The background level is indicated by the gray line.
Determination of Protein Content
Whole-muscle cross-sections of EDL and soleus were used to determine the protein content by a histochemical method using naphthol Yellow S as described by Tas and James (1981). In addition, EDL and soleus muscle samples were homogenized in 20 volumes of ice-cold bidistilled water and used for the colorimetric protein detection method based on bicinchoninic acid (BCA-kit; Pierce, Rockford, IL). Bovine serum albumin was used as a standard for the calibration curve.
Results
The ISH analysis using the Types I, IIA, IIX, and IIB isoform-specific MHC probe, is shown in Figure 2. A strong hybridization signal is observed in slow Type I fibers compared to the fast Type II fibers. Furthermore, in the group of Type II fibers the hybridization signal is highest in Type IIA fibers and decreases towards Type IIX and IIB fibers. Hybridization of the sections to another MHC probe than its nominal one, e.g., fiber Type I to the probe for MHC IIA, IIX, or IIB did not reveal any signal (not shown).
A first conclusion from these observations was that the amount of MHC mRNA differs in the separate fiber types. This conclusion may imply a difference in MHC mRNA content in fibers that have equal total RNA content. On the other hand, the observation can equally well be explained by assuming a difference in total RNA content in fibers that have equal amounts of MHC mRNA per μg of total RNA. Obviously, a mixture of both possibilities can be envisioned as well. To test whether different amounts of MHC mRNA per μg of total RNA are present in slow and fast fibers, a Northern blot analysis was carried out. Total RNA was isolated from soleus (more than 90% Type I fibers) and EDL (more than 90% Type II fibers) muscle. Equal amounts of RNA were fractionized on size by gel electrophoresis, blotted onto a nitrocellulose membrane, and allowed to hybridize to a 32PTabeled probe specific for the ATP binding site of human β-MHC mRNA. This domain is highly conserved throughout different myosin isoforms and therefore can be used as a marker for the total MHC mRNA content. Figure 3 shows that the intensity of hybridization was similar with both RNA preparations, indicating that equal amounts of MHC mRNA are present per μg of total RNA.
Expression of Types I, IIA, IIX and IIB MHC mRNA in histochemically characterized single rat muscle fibers by ISH with the isoform-specific MHC probe. A strong hybridization signal can be seen in the Type I fiber, and the signal decreases in the fast fiber types.
The same probe was then used to test whether it could disclose differences in MHC mRNA content per fiber. To get a first indication of whether such differences would exist, an ISH experiment was performed. Whole-muscle cross-sections from soleus and EDL muscle were hybridized with the probe to the MHC ATP binding site.
A high density of grains can be seen in Type I fibers from soleus muscle (Figure 4A) compared to the level in the EDL muscle (Figure 4C), which hybridizes more weakly and heterogeneously. This heterogeneity is due to the presence of some Type I fibers that display high levels of expression, as revealed by comparing the serial sections used for ISH and mATPase histochemistry (Figures 4B, 4D, and 4E). It is clear that differences in expression level between Type I and IIA fibers in soleus muscle (Figure 4A) cannot be reliably discriminated because the few IIA fibers present in this muscle are all surrounded by Type I fibers. In the EDL muscle, a distinction can be made because the Type IIA fiber is surrounded by Type IIB fibers, which show a very low expression level (Figure 4C). In agreement with the hybridization data shown in Figure 2, these data support the hypothesis that MHC mRNA content is high in Type I fibers and low in fast fibers. In combination with the Northern blot data, this implies that total RNA content differs in the separate fiber types. This implication was confirmed by ISH on sections from the EDL and soleus muscle in which a cRNA probe for 28S rRNA was used as a marker for total RNA. The results observed in both muscle sections are similar to the data described with the MHC ATP binding site probe (not shown). To see whether more general genes that are not tissue specific show similar pronounced differences among the distinct fiber types, a cRNA probe for elongation factor-1α was used. Whole-muscle cross-sections from EDL and soleus muscle were hybridized with this probe and a high density of grains was observed in the cross-section of the soleus muscle and in the few Type I fibers present in the EDL muscle. However, we were not able to show clear differences between the distinct fast fiber types present within the EDL muscle because of the overall low level of expression (not shown).
Northern blot analysis using the MHC ATP binding site cRNA probe in rat soleus and EDL muscle. Total RNA was isolated from both muscles and 10 μg of total RNA was applied in each lane. Equal amounts of MHC mRNA were observed in Lane 1 (soleus muscle) compared to Lane 2 (EDL muscle). Ribosomal RNA staining was used as the internal standard for the amount of RNA loaded on the gel.
To test directly whether total RNA content differs in the distinct fiber types, a quantitative analysis of total RNA was performed in single muscle fibers by ISH and densitometry in distance zones. As a marker for the amount of total RNA, we used a cRNA probe for 28S rRNA. Fibers were carefully classified as Types I, IIA, IIX, and IIB according to the combined staining patterns for the mATPase methods. As shown in Figure 5, only fibers expressing Type I MHC stain after preincubation at pH 4.4. Fixation of the fibers followed by preincubation in alkaline media (pH 10.4) differentiated three fiber populations: Type I fibers (negative), Types IIA and IIX (black), and Type IIB fibers (gray). The double preincubation method allows the distinction of IIX fibers because this method stains exclusively fibers expressing IIX MHC. The fourth row in Figure 5 shows the hybridization signal in examples of these four distinct fiber types using the 28S rRNA probe. It is clear that Type I fibers have the strongest expression of 28S rRNA. Among the fast fiber types the expression is highest in Type IIA fibers and decreases in IIX and IIB fibers. Both controls, ISH with the sense RNA probe and pretreatment with RNase, resulted in nondetectable staining.
ISH using the MHC ATP binding site cRNA probe in whole-muscle cross-sections of soleus (A) and EDL (C) muscle. Type I fibers in soleus and EDL muscle show a much higher expression level compared to the level in the distinct Type II fibers. Fiber typing was performed by staining transverse serial sections for mATPase after preincubations at (B) pH 4.4, (D) pH 4.6, and (E) pH 10.4 following formaldehyde fixation.
Quantification of the density of grains (IOD/area; mean ± SEM) is summarized in Table 1 and shows a systematic decrease in the amount of rRNA per amount of tissue in the order i>IIA>IIX>IIB. Type I fibers have twofold higher expression level compared to Type IIA fibers, and when Type I fibers are compared with the fastest Type IIB fibers a five- to sixfold higher expression level is observed in the Type I fibers. All differences observed were statistically significant (p<0.05).
Total muscle fiber RNA content assessed by 28S rRNA in situ hybridization and orcinol assay! a
For the in situ hybridization, IOD values are corrected for fiber cross-sectional area and are presented as mean ± SEM.
Significantly higher compared to Types IIA, IIX, and IIB fibers;
significantly higher compared to Types IIX and IIIB fibers;
significantly higher compared to Type IIB fibers (p<0.05). For the orcinol, assay total RNA content is expressed as μg/mg dry weight. For both methods, the relative total RNA content (± SEM) for Types IIA, IIX, and IIB fibers was determined with the mean value for the Type I fibers set to 100.
To verify these results with an independent procedure, the amount of total RNA in the distinct fiber types was determined with the orcinol method (see Table 1). The highest total RNA content is present in Type I fibers. Among the fast fiber types, total RNA content is highest in IIA fibers and decreases towards Types IIX and IIB fibers, thus corroborating the data of the 28S rRNA quantification. Furthermore, when the total RNA determination for the IIA, IIX, and IIB fiber types is expressed in relation to the value for the Type I fibers, very similar values to those demonstrated by the 28S rRNA quantification are revealed (Table 1). The relative RNA content of the orcinol method is well within the 95% confidence interval of the values based on the ISH.
To assess whether the differences in total RNA level between slow fibers and fast fibers are reflected in similar differences in total protein content, we performed the histochemical protein content analysis on sections. No differences among fiber types were detected. Subsequently, soleus (predominantly slow fibers) and EDL (predominantly fast fibers) muscles were homogenized and used to determine total protein content by the colorimetric procedure. Again, no differences were found using this method (soleus 20.5 ± 1.4 μg/ml; EDL 20.9 ± 1.7 μg/ml).
Discussion
The primary focus of this study was to determine whether the amount of RNA varies among the distinct fiber types (I, IIA, IIX, and IIB) in adult rat skeletal muscle. To our knowledge, this is the first report presenting data at the single-fiber level that quantifies a systematic difference among the four distinct fiber types found in adult rat skeletal muscle. Using two independent methods (in situ hybridization and the biochemical orcinol assay), Type I fibers show a twofold higher total RNA content compared to Type IIA fibers. When we compared Type I fibers with the fastest Type IIB fibers, there was a five- to sixfold higher total RNA content. In addition, Northern blot analysis demonstrated that the amount of MHC mRNA per μg of total RNA is similar in the separate fiber types. Therefore, the MHC mRNA content is different in the distinct fibers. Similar differences were also observed between Type I and Type II fibers for a general housekeeping gene (i.e., EF-1α). Consequently, the contribution of a distinct fiber type to the total amount of RNA must be taken into account in studies of muscle gene expression. For example, if we consider the red area of the MG, which is composed of 10% Type I, 10% Type IIA, 45% Type IIX, and 35% Type IIB fibers (de Ruiter et al. 1996), the few Type I fibers will contribute almost 30% of the total muscle RNA content. To date, mRNA measurements have not been corrected for these differences, although it has previously been shown that the total RNA and polyadenylated RNA concentration is higher in whole-muscle preparations from slow muscles (e.g., soleus muscle) than from mixed or fast muscles (e.g., MG and EDL muscle). The reason may be the variability in muscle total RNA concentrations presented thus far (e.g., Goldberg 1967; Steffen and Musacchia 1984; Kurowski et al. 1987; Hood and Simoneau 1989; Choo et al. 1992).
ISH and histochemical analysis on serial cryostat sections of gelatin-embedded rat single-muscle fibers. ISH was performed using the mouse 28S rRNA probe. Histochemical classification of fiber types was achieved after fixation and preincubation at pH 4.4, 10.4, and sequential preincubation in alkaline (pH 10.4) and acidic (pH 4.5) media. The measured density distribution (IOD) of each fiber is presented in the bottom row. The white line in each graph represents the border of the cell. For the labels on the axes, the reader is referred to Figure 1F.
Because the majority of cellular RNA (80-85%) is ribosomal, the rRNA content has been used as an index of the protein synthetic capacity of the fiber (Millward et al. 1973). It has been shown at the whole-muscle level that a slow-twitch muscle has a two- to threefold higher protein synthesis rate than a fast-twitch muscle, which correlates well with the higher total RNA content (e.g., Goldberg 1967; Flaim et al. 1980; Hood and Terjung 1987). However, no data are presently available with regard to protein synthetic capacity or protein content at the single-fiber level. Our single-fiber data show systematic differences in total RNA content for each fiber type and therefore suggest that protein synthetic capacity and protein content would be highest in Type I fibers and would decrease towards the faster fiber types. To this end, we determined protein content in whole-muscle preparations from soleus and EDL muscle. Surprisingly, no differences were found with either method. This implies that no major differences in MHC protein content exist among different fiber types despite large differences in MHC mRNA content. This notion is underscored by single-fiber analysis using SDS-PAGE (Sant'Ana Pereira et al. 1995a). Assuming that the rRNA content reflects the number of ribosomes, slow fibers have a sixfold higher capacity to synthesize protein. Because total protein content is not higher, either the potential capacity is not used or the protein turnover is higher in Type I fibers.
Because protein synthesis is a major determinant of energy turnover in the cell, the oxidative metabolic capability of a fiber can be expected to be related to the total RNA content. Among the fast fiber types, a good relationship exists between oxidative capacity and total RNA content. However, the correlation is less clear if we compare Type I vs Type II fibers. The highest total RNA content was observed in Type I fibers. In rodents, however, succinate dehydrogenase (SDH) activity, which has been used as a marker for oxidative capability, is slightly higher in Type IIA compared to Type I fibers (Schiaffino et al. 1989b; de Ruiter et al. 1995; and our unpublished observations). Similarly, mitochondrial content has also been shown to be higher in Type IIA compared to Type I fibers. The explanation for this apparent discrepancy may lie in the relative balance between the energy turnover required for protein synthesis and that required for the generation of mechanical output by the different fiber types. Type I fibers, which have slow crossbridge detachment rates suited to a postural role, will have a markedly lower energy turnover compared with faster fiber types, which are recruited in locomotive activities requiring active shortening and power production (de Ruiter et al. 1996). Clearly, the balance of total postural to locomotive activity may vary among species and hence may change the total demand for oxidative phosporylation in relation to that proportion attributable to protein synthesis.
The mechanism by which a higher total RNA content is achieved can also be related to the number of myonuclei per fiber. Several studies demonstrated that the amount of myonuclei per mm of myofiber is greater in Type I fibers compared to Type II fibers (e.g., Schmalbruch and Hellhammer 1977; Tseng et al. 1994; Allen et al. 1996), although none of these studies made further subdivision within the Type II fibers. If a correlation with myonuclei number holds true, rRNA content would be dependent on gene doses rather than on transcriptional rate or rRNA stability. Further analysis needs to be done to test this possibility.
A second issue that deserves attention is the intracellular localization of the RNA. Our analysis of the distribution of the silver grains reveals a symmetric distribution of grains, with a peak density just beneath the sarcolemma (Figures 1F and 5). Given the range of the β-particles of 35S (10 μm), we conclude that in adult rat muscle fibers the majority of RNA is subsarcolemmally localized. This conclusion is in line with the localization of the ribosomes (e.g., Home and Hesketh 1990) and with ISH studies on whole-muscle cross-sections (e.g., Dix and Eisenberg 1988; Aigner and Pette 1990). The symmetric distribution of grains does not support an intermyofibrillar localization of RNA as suggested by a study of Hesketh et al. (1991), who showed by ISH on longitudinal and transverse sections of soleus muscle of young rats that the MHC mRNA was distributed throughout the myofibers. They found greater density of grains in the subsarcolemmal regions of the fibers but there was also a considerable number of grains in the core myofibrillar region of the fibers. Their distribution of mRNA might be a characteristic of immature fibers, and it is conceivable that the subsarcolemmal localization of RNA in these fibers is acquired with their maturation.
In summary, this study demonstrates for the first time that the amount of total RNA per amount of tissue varies systematically among Types I, IIA, IIX and IIB fibers in rat hindlimb muscles. ISH on sections of single fibers followed by quantification and independent biochemical method both demonstrate that the amount of total RNA decreases significantly from i>IIA>IIX>IIB; slow Type I fibers show a five- to sixfold higher expression level compared to the fastest Type IIB fibers. The MHC mRNA content per μg of total RNA is shown to be comparable in all fiber types. Therefore, Type I fibers also contain five- to sixfold more MHC mRNA per μg of tissue compared to the fastest Type IIB fibers. This implies that proper mRNA quantification in slow vs fast muscle fibers and also among fast Type II fibers requires correction for the total amount of RNA per fiber type.
Footnotes
Acknowledgments
Supported in part by NWO Grant number 902-16-219 (to DF).
We gratefully acknowledge Piet de Boer for help with in situ hybridization, J. Hagoort for image aquisition, Cars E. Gravemeijer and Cees J. Hersbach for excellent photography, and Dr Stefano Schiaffino and Dr Margaret Buckingham for kindly providing the IIA, IIX, and IIB MHC cDNAs.