Membrane-associated carbonic anhydrase IV in ... - Springer Link

4 downloads 0 Views 4MB Size Report
&p.1:Abstract Carbonic anhydrase IV (CA IV) was exam- ined by light microscopy and electron microscopy in rat soleus muscle. Semithin sections of aldehyde- ...
Histochem Cell Biol (1996) 106:405–411

© Springer-Verlag 1996

O R I G I N A L PA P E R

&roles: Brigitte Decker · Susanne Sender · Gerolf Gros

Membrane-associated carbonic anhydrase IV in skeletal muscle: subcellular localization

&misc: Accepted: 17 May 1996

&Abstract p.1: Carbonic anhydrase IV (CA IV) was examined by light microscopy and electron microscopy in rat soleus muscle. Semithin sections of aldehyde-fixed Epon-embedded muscle were stained with rabbit anti-rat lung CA IV and the avidin-biotin-peroxidase complex. With this technique, capillaries and sarcolemma showed positive CA IV staining. For electron microscopy, rat soleus specimens were aldehyde-fixed, with or without subsequent osmication, and embedded in Epon. Ultrathin sections were immunostained with anti-rat lung CA IV/immunogold. Omitting osmium allowed ample antigen-antibody reactions but could not prevent the release of glycosylphosphatidylinositol-anchored CA IV from the membranes, which led to apparent background staining. Postosmication significantly reduced tissue antigenicity but kept the antigen bound to the membranes and thus allowed a very precise localization of CA IV. By electron microscopy, membrane-bound CA IV is found to be associated with capillary endothelium, sarcolemma, and sarcoplasmic reticulum (SR). Conceivably, the presence of SR staining in ultrathin sections and its absence in semithin sections reflect a problem of accessibility of the antigenic sites.&y b :d

Introduction Apart from the cytosolic carbonic anhydrase (CA) isozymes CA II and CA III, mammalian skeletal muscle fibers have been shown to possess at least one membranebound CA form, CA IV (Gros and Dodgson 1988; Waheed et al. 1992). A number of studies indicated that a B. Decker (✉) Abteilung Zellbiologie und Elektronenmikroskopie, Medizinische Hochschule Hannover, D-30623 Hannover, Germany Tel. +49-511-5323912; Fax +49-511-5323903 S. Sender · G. Gros Vegetative Physiologie, Zentrum Physiologie, Medizinische Hochschule Hannover, D-30623 Hannover, Germany&kl / f o n - b:c

membrane-bound CA is associated with the sarcolemma (SL) and the sarcoplasmic reticulum (SR) (Dermietzel et al. 1985; Geers et al. 1985; Bruns et al. 1986; Wetzel et al. 1990). However, no immunohistochemical evidence for a CA IV at these locations has been reported so far. In a recent cryosection study using antisera against rat lung and human lung CA IV, we demonstrated the existence of membrane-bound CA IV in capillary endothelium of rat and human skeletal muscles (Sender et al. 1994). In that study, the muscle fibers themselves showed no detectable staining. Considering the strong evidence for CA in SL and SR, it was the aim of the present study to screen muscle tissue for membrane-bound CA IV at the subcellular level. For this purpose, semithin and ultrathin sections of rat soleus were immunostained for CA IV.

Materials and methods Preparation of rat soleus Small specimens of the muscle were fixed for 4–6 h with: (a) 1% glutaraldehyde/2% paraformaldehyde in 0.1 M sodium cacodylateHCl buffer, pH 7.3, without postosmication; or (b) 2.5% glutaraldehyde/2% paraformaldehyde in 0.1 M sodium cacodylate-HCl buffer, pH 7.3, followed by postfixing for 1 h with 1% OsO4 dissolved in the same buffer. After dehydration in a graded series of ethanol, the samples were embedded in Epon 812 (Serva, Heidelberg, Germany). Immunohistochemistry Light microscopy Semithin sections of specimens fixed without postosmication were cut at 1 µm with a LKB 4800-III ultrotome and mounted on polyL-lysine-coated slides (Sigma, Deisenhofen, Germany). After removal of the epoxy resin by sodium methoxide (Mayor et al. 1961), the sections were incubated with polyclonal rabbit anti-rat lung CA IV (diluted 1:100, 1:200, 1:300 or 1:400) for 24 h at 4° C, followed by incubation with the second antibody, biotin-labeled goat anti-rabbit IgG (dilution 1:200; Dianova, Hamburg, Germany) for 30 min. Phosphate-buffered saline (PBS, 0.2 M, pH 7.3) was used as diluent for the antisera and as rinsing solution. The sections were then incubated for 30 min with a preformed com-

406 plex of biotin-peroxidase (0.7 µg/ml)/streptavidin (5 µg/ml) (Dianova) diluted in PBS. The antibody-binding sites were visualized by incubating the sections for 10 min in a solution containing 0.5 mg/ml diaminobenzidine tetrahydrochloride (DAB; Sigma) in 0.05 M TRIS-HCl, pH 7.6 with 0.002% H 2O2. Controls were carried out by incubating the sections with rabbit non-immune serum instead of antibody at identical dilutions. Electron microscopy Ultrathin sections, about 80-nm-thick, were collected on nickel grids and immunostained by the immunogold technique of De Mey (1983). In brief, the sections, after etching with a saturated solution of sodium metaperiodate (30 min), were incubated overnight at 4° C with a solution of 0.1% BSA/0.05 M TRIS-HCl, pH 7.6, and rabbit anti-rat lung CA IV (diluted 1:150, 1:200 or 1:300). The sections were then rinsed in 0.1% BSA/0.05 M TRIS-HCl and incubated for 1 h with gold-labelled anti-rabbit IgG (diluted 1:10 or 1:20; Amersham Buchler, Braunschweig, Germany). The size of the gold particles was 10 nm. Following a second wash in 0.1% BSA/0.05 M TRIS-HCl, sections were stained with a saturated aqueous solution of uranyl acetate and examined in a Siemens Elmiskop 101 electron microscope. Control sections were either incubated with rabbit non-immune serum (Sigma), using the same dilutions as for the antiserum, or with antigen-blocked antiserum, as described by Sender et al. (1994).

Results Light microscopy The endothelial cells of the capillaries and the SL of the muscle fibers were stained after incubation of semithin sections with anti-CA IV/avidin-biotin-peroxidase complex (ABC) (Fig. 1a). The sarcoplasm did not show CA IV immunoreactivity. The texture visible inside the muscle fibers is rather due to the etching procedure, since it was also seen in control sections (Fig. 1b) which showed no immunoreactivity. The immunostaining seen in Fig. 1a is therefore considered to be specific and indicates the association of CA IV with capillary endothelial cells and SL. Electron microscopy In the absence of osmication, the ultrastructure of the rat soleus muscle was reasonably well preserved (Figs. 2, 3). Incubation of these sections with anti-CA IV/anti-rabFig. 1a, b Semithin sections from rat soleus, fixed with 1% glutaraldehyde/2% paraformaldehyde. a Immunostaining with anticarbonic anhydrase (CA) IV 1:400/avidin-biotin-peroxidase complex (ABC). Capillaries (indicated by arrows) and sarcolemma (SL) show a positive reaction. b Incubation with rabbit non-immune serum 1:400/ABC. Staining is reduced to background level. Bars, 40 µm&.i c/ :gf

407

Fig. 2a–c Ultrathin sections from rat soleus, fixation as in Fig. 1, sections treated with anti-CA IV 1:300/immunogold. In a, arrow indicates positive reaction of capillary endothelium, arrowhead points to positive staining of SL. In b, capillary endothelium (EN) with a positive reaction. Some weak staining appears in the capillary lumen and on the basal lamina (BL). c Incubation with antigen-blocked antiserum, a very weak unspecific staining can be seen. Bars a, 0.1 µm, b, 0.2 µm, c, 0.4 µm&.i c/ :gf

bit IgG immunogold showed patches of gold particles in different locations (Figs. 2a, b, 3a). The gold particles were predominantly found in those locations which showed CA IV immunoreactivity in semithin sections, as depicted, for example, in Fig. 1a. The strongest CA IV immunoreactivity was seen in the capillary endothelium (Fig. 2a, b) and in the SL (Fig. 2a). A few, mostly single,

408

409

Fig. 4a, b Ultrathin sections of rat soleus, fixed with 2.5% glutaraldehyde/2% paraformaldehyde, postfixed in OsO4, and immunostained with anti-CA IV 1:200/immunogold. In a, immunogold particles are seen at the luminal membrane of capillary endothelium (arrowhead) and at the SL (arrows). In b, immunogold particles are seen in association with SR membranes, located between the myofibrils. Bars, 0.2 µm&.i c/ :gf

gold particles were occasionally present in the capillary lumen and outside the endothelial basal lamina (Fig. 2a, b). Clusters of gold particles were also seen between the myofibrils (Fig. 3a), where the SR is known to form its three-dimensional network. Only a few, mostly single, particles were seen within the myofibrils. Control sections incubated with antigen-blocked antiserum showed only very few scattered gold particles (Figs. 2c; 3b). We conclude, therefore, that the immunoreactions shown in Figs. 2a, b and 3a are indeed specific and that CA IV is associated with capillary endothelium, SL, and SR. Specimens fixed with aldehydes and postfixed with 1% OsO4 prior to immunostaining resulted in better preservation of muscle fiber ultrastructure (Fig. 4). It

also resulted in a precise localization of antigenic sites with little background staining. In contrast to the sections of muscle tissue which had not been osmicated, no diffuse distribution of single immunogold particles was observed (compare Figs. 2a, b, 3a with Fig. 4a). On the other hand, the absolute number of reactive antigenic sites has drastically decreased as a consequence of the osmium treatment (and higher glutaraldehyde concentration). Anti-CA IV/immunogold staining was seen at the luminal membrane of capillary endothelial cells and at the SL (Fig. 4a). Furthermore, some T tubules were also stained with immunogold particles (not shown). Clusters of immunogold particles were again observed in distinct areas between the myofibrils in close association with the SR membranes (Fig. 4b). Mitochondrial membranes remained unstained. No staining was seen in control sections that were incubated with rabbit non-immune serum/immunogold (not shown). Thus, the staining shown in Fig. 4 confirms the results shown in Figs. 2 and 3 but, due to postosmication of the tissue, allows a more exact localization of CA IV.

Discussion Fig. 3a, b Ultrathin sections from rat soleus, fixation as in Fig. 1. In a, incubation with anti-CA IV 1:300/immunogold. Several areas around the myofibrils (indicated by arrows) show accumulation of immunogold particles. The stained structures very likely represent sarcoplasmic reticulum (SR). In b, incubation with antigen-blocked antiserum 1:300/immunogold. Almost no unspecific staining can be seen. Bars, 0.2 µm&.i c/ :gf

The results presented in this study were obtained using aldehyde fixatives with and without postosmication. Omitting OsO4 led to a remarkable antigen-antibody reaction, which was greatly reduced after postosmication. OsO4 is known to destroy tissue antigenicity to a signifi-

410

cant extent (Polak and Priestley 1992). Nevertheless, postosmication was useful to obtain the exact localization of the reaction product, as CA IV is a glycosylphosphatidylinositol-anchored antigen (Zhu and Sly 1990) which, without postosmication, is probably released to an appreciable extent from the lipid phase of the membranes during tissue processing with organic solvents and diffuses into neighboring structures. OsO4, by fixing proteins and lipids, keeps both at their original sites, preventing movement (and/or eventual loss) of membrane components (Romeis 1989; Polak and Priestley 1992). CA IV in capillary endothelium The staining of the capillary endothelium both in the semithin sections and at the electron microscope level supports previous results obtained with the indirect immunofluorescence method on cryosections of skeletal muscle (Sender et al. 1994). Figure 4a shows that CA IV is bound to the luminal membrane of the endothelial cell, similar to the binding which has been shown for brain capillaries by Ghandour et al. (1992). CA IV in the SL The SL gave a positive reaction for CA IV on the light as well as on the electron microscope level. This result extends the findings of our previous immunofluorescence study in which SL staining could not be demonstrated on 5- to 7-µm-thick cryosections (Sender et al. 1994). This difference may be explained by the greater sensitivity of the ABC technique employed in the present investigation. Also, the embedding procedure used for the semithin sections causes considerably greater cell shrinkage than the procedure used for the cryosections. This leads to a widening of the intercellular space and therefore may facilitate the access of antibodies to the outer surface of the SL. The association of a CA with the SL of skeletal muscle fibers has already been inferred. Ridderstråle (1979), Lönnerholm (1980), and Riley et al. (1982) showed sarcolemmal staining with the modified Hansson method, whereas Dermietzel et al. (1985) and Geers et al. (1985) achieved a positive reaction with dansylsulfonamide (DNSA), a fluorescent CA inhibitor. Indirect evidence for the existence of a sarcolemmal CA was obtained in perfusion experiments by Geers et al. (1985) as well as by De Hemptinne et al. (1987), whereas Wetzel and Gros (1990) obtained biochemical evidence by measuring CA activity in sarcolemmal vesicles. Further experiments identified this activity as CA IV (Waheed et al. 1992). A sarcolemmal CA could be important for the rapid buffering of fixed acid H+ in the interstitial space during muscle exercise. CA IV in the SR In the present study, CA IV could be demonstrated in the SR on ultrathin but not on semithin sections. The prob-

lem of staining the SR has already been discussed in a previous study (Sender et al. 1994) showing that the SR did not react with anti-CA IV on muscle cryosections. Nevertheless, several authors did obtain a reaction with the SR on cryosections with DNSA (Dermietzel et al. 1985; Geers et al. 1985; Bruns et al. 1986). This strongly indicates a problem of accessibility of CA IV to the histochemical reagents. As a major part of the SR CA is oriented towards the SR interior (Geers et al. 1991; Gros et al. 1993), probably a lipophilic, membrane-permeable substance such as DNSA is required to gain access to the enzyme, while macromolecules such as antibodies cannot enter the intact SR structure. In the sections used for the ABC and the immunofluorescence techniques (about 1-µm and 5- to 7-µm-thick, respectively), the majority of the SR is presumably present as an intact, closed tubular network, into which antibodies cannot penetrate. In ultrathin sections, the section thickness of about 0.08 µm is less than the thickness of longitudinal tubules. Therefore, access of macromolecules to the interior surface of the SR membrane is possible and immunohistochemical staining becomes apparent at the electron microscope level. The same reasoning holds for the few T tubules that have been found stained in the present study. The existence of CA associated with the SR is supported by biochemical and physiological results. CA activity could be measured in SR vesicle preparations (Wetzel and Gros 1990; Bruns et al. 1986). Also, Wetzel et al. (1990) measured Ca++ transients in rat soleus fibers that suggest a role of SR CA in excitation-contraction coupling. &Acknowledgements p.2: We would like to thank Professor William S. Sly, St. Louis, USA, for providing us with the antiserum against rat lung CA IV. We also thank Dr. A. Jörns, Abteilung Allgemeine Anatomie, Medizinische Hochschule Hannover, for valuable advice in immunohistochemical techniques. The excellent technical assistance of Mrs. I. Albers and the expert photographic assistance of Mrs. A. Hundt is gratefully acknowledged. This work was supported by the Deutsche Forschungsgemeinschaft, grant Gr 489/11.

References Bruns W, Dermietzel R, Gros G (1986) Carbonic anhydrase in the sarcoplasmic reticulum of rabbit skeletal muscle. J Physiol 371:351–364 De Hemptinne A, Marrannes R, Vanheel B (1987) Surface pH and the control of intracellular pH in cardiac and skeletal muscle. Can J Physiol Pharmacol 65:970–977 De Mey J (1983) Colloidal gold probes in immunocytochemistry. In: Polak JM, Van Norden S (eds) Immunocytochemistry, practical applications in pathology and biology. Wright-PSG, Bristol, pp 55–68 Dermietzel R, Leibstein A, Siffert W, Zamboglou N, Gros G (1985) A fast screening method for histochemical localization of carbonic anhydrase. Application to kidney, skeletal muscle, and thrombocytes. J Histochem Cytochem 33:93–98 Geers C, Gros G, Gärtner A (1985) Extracellular carbonic anhydrase of skeletal muscle associated with the sarcolemma. J Appl Physiol 59:548–558 Geers C, Wetzel P, Gros G (1991) Is carbonic anhydrase required for contraction of skeletal muscle? News Physiol Sci 6:78–81 Ghandour MS, Langley OK, Zhu XL, Waheed A, Sly WS (1992) Carbonic anhydrase IV on brain capillary endothelial cells: a

411 marker associated with the blood-brain barrier. Proc Natl Acad Sci USA 89:6823–6827 Gros G, Dodgson S (1988) Velocity of CO2 exchange in muscle and liver. Annu Rev Physiol 50:669–694 Gros G, Wetzel P, Geers G, Dodgson S, Forster RE, Sender S (1993) Functional significance of carbonic anhydrase in skeletal muscle. In: Scheid P (ed) Funktionsanalyse biologischer Systeme, vol 23. Gustav Fischer, Stuttgart, pp 189–196 Lönnerholm G (1980) Carbonic anhydrase in rat liver and rabbit skeletal muscle; further evidence for the specificity of the histochemical cobalt-phosphate method of Hansson. J Histochem Cytochem 28:427–433 Mayor HD, Hampton JC, Rosario B (1961) A simple method for removing the resin from epoxy-embedded tissue. J Cell Biol 9:909–910 Polak JM, Priestley JV (1992) Electron microscopic immunocytochemistry: principles and practice. Oxford University Press, Oxford Ridderstråle Y (1979) Observations on the localization of carbonic anhydrase in muscle. Acta Physiol Scand 106:239–240

Riley DA, Ellis S, Bain J (1982) Carbonic anhydrase activity in skeletal muscle fiber types, axons, spindles and capillaries of rat soleus and extensor digitorum longus muscles. J Histochem Cytochem 30:1275–1288 Romeis B (1989) Mikroskopische Technik, 17th edn. Urban and Schwarzenberg, Munich Sender S, Gros G, Waheed A, Hageman GS, Sly WS (1994) Immunohistochemical localization of carbonic anhydrase IV in capillaries of rat and human skeletal muscle. J Histochem Cytochem 42:1229–1236 Waheed A, Zhu XL, Sly WS, Wetzel P, Gros G (1992) Rat skeletal muscle membrane associated carbonic anhydrase is 39-kDa, glycosylated, GPI-anchored CA IV. Arch Biochem Biophys 294:550–556 Wetzel P, Gros G (1990) Sarcolemmal carbonic anhydrase in red and white rabbit skeletal muscle. Arch Biochem Biophys 279:345–354 Wetzel P, Liebner T, Gros G (1990) Carbonic anhydrase inhibition and calcium transients in soleus fibers. FEBS Lett 267:66–70 Zhu XL, Sly WS (1990) Carbonic anhydrase IV from human lung. J Biol Chem 265:8795–8801