THE JOURNAL OF COMPARATIVE NEUROLOGY 326273-282 (1992)
Immunofluorescent Analysis of Creatine Kinase in Cultured Astrocytes by Conventional and Confocal Microscopy: A Nuclear Localization PATRICIA MANOS AND JOHN EDMOND Department of Biological Chemistry and Mental Retardation Research Center, UCLA School of Medicine, Los Angeles, California 90024
ABSTRACT The subcellular localization of creatine kinase (CK) was examined in primary cultures of astrocytes with immunofluorescent labeling methods and detection by both standard fluorescence microscopy and confocal laser-scanning microscopy. With conventional microscopy, the pattern of CK staining was uniform throughout the cell cytoplasm and appeared to stain the nuclear region intensely. Staining of CK in the nuclear region co-localized with the DNAspecific Hoechst nuclear stain. CK produced a diffuse cytoplasmic staining pattern that was different from the staining pattern produced by the cytoskeletal proteins glial fibrillary acidic protein and tubulin, both of which showed a filamentous cytoskeletal network that excluded the nucleus. To examine the structural details of CK in the nuclear region, serial optical sections were taken through the cell monolayer with a confocal microscope. The cells were immunostained for CK, and the CK-staining pattern was compared with the staining pattern produced by propidium iodide, which is specific for DNA in RNase-treated samples and stains total nucleic acid in untreated samples. CK staining was present within the nucleus in each section taken through the monolayer. The nucleolus did not stain for CK. The pattern of CK staining in the nucleus (and cytoplasm) was distinctly different from the staining pattern of either DNA or total nucleic acid. Nuclear CK appeared to have a granular, particulate pattern, which is suggestive of a nucleoplasmic distribution. o 1992 Wiley-Liss, Inc. Key words: creatine phosphokinase, cerebral energy metabolism, glia
The mammalian central nervous system has a high rate of oxidative metabolism and is believed to use, a large portion of the total body energy requirement (Hawkins, ’85).Creatine, creatine phosphate, and creatine kinase (CK, EC 2.7.3.2)play an important role in brain energy homeostasis, in part by maintaining cellular ATP levels. CK catalyses the reversible transfer of a high energy phosphate between ATP and creatine phosphate and is high in tissues with high energy requirements like brain, skeletal muscle, and heart. CK is a dimeric enzyme with four different molecular forms (for reviews, see: Watts, ’73; Kenyon and Reed, ’83; Bessman and Carpenter, ’85; Wallimann et al., ’92). One isoenzyme is located on the inner mitochondrial membrane and functions in respiratory control. The other three isoforms are believed to be primarily cytoplasmic and are expressed in a tissue-specific fashion. The “B” subunit type is the predominant form in brain, whereas the “M” subunit type is the predominant form in skeletal muscle; the heterodimer MB form is characteristic for heart. Cytoplasmic CK is believed to provide a sustained, local, and
o 1992 WILEY-LISS, INC.
high concentration of ATP at cellular sites of high energy utilization. The CK reaction has been described as a near-equilibrium reaction that functions as an energy reservoir, or as a buffer for ATP levels (Watts, ’73; Kenyon and Reed, ’83; Meyer et al., ’84). Alternatively, arguments based on the metabolic compartmentation of adenine nucleotides and the CK isoenzymes have led to the idea of an energy “shuttle” (Bessman and Geiger, ’81;Wallimann and Eppenberger, ’90). High energy phosphate in the form of creatine phosphate is targeted to specific sites in the cell where cytoplasmic CK is associated with an ATPase. Cellular ATPases are unable to use the high energy phosphate in creatine phosphate unless CK is present. Creatine is then rephosphorylated by the mitochondrial isoenzyme. The creatine-creatine phosphate energy shuttle has been well Accepted August 14,1992 Address reprint requests to Dr. Patricia Manos, Department of Biological Chemistry, 33-257 Center for the Health Sciences, UCLA School of Medicine, Los Angeles, CA 90024.
P. MANOS AND J. EDMOND
274 described in muscle (Bessman and Geiger, '81; Meyer et al., '84; Bessman and Carpenter, '85; Wallimann and Eppenberger, '90; Wallimann et al., '92) and sperm (Tombes and Shapiro, '85). CK activity increases rapidly in postnatal brain development and is present in homogeneous populations of cultured neurons, oligodendrocytes, and astrocytes (Manos et al., '91). The metabolic activity of astrocytes is high and is thought to be comparable to that of neurons (Hertz, '81). Many metabolic pathways, especially those related to energy metabolism, are unique to astrocytes. Astrocytes function in glutamate uptake and ammonia fixation (Martinez-Hernandez et al., '77; Hertz, '81). Astrocytes have the capacity to oxidize fatty acids (Edmond et al., '87) as well as synthesize ketone bodies (Austead et al., '91) in support of the energy requirements in developing brain. Because of the importance of astrocytes in cerebral energy metabolism, we were interested in examining the subcellular localization of CK in cultured astrocytes. With conventional fluorescent microscopy, immunoreactive CK was found to be dispersed throughout the cytoplasm and to stain the nuclear region of the cell intensely. The staining pattern of CK was distinct from the staining pattern of cytoskeletal proteins and co-localized with the Hoechst stain. Serial optical sections taken through the cell monolayer with a confocal microscope indicate that CK is present uniformly throughout the nucleus, does not co-localize with either DNA or RNA, and has a punctate, or granular, pattern that indicates it is located in the nucleoplasm.
MATERIAL AND METHODS Rats were purchased from Bantam and Kingman; cell culture medium was from GIBCO; Calf Serum (Lot #2 151682) was from Hyclone Laboratories. The antibodies were obtained commercially from the following sources: rabbit antihuman CK-BB (Calbiochem); goat antihuman CK-BB (Chemicon); mouse anti-GFAP (Boheringer Mannheim); mouse antitubulin (Boheringer Mannheim). The secondary antibodies, purchased from Jackson Laboratories, were affinity purified IgG with minimal cross reactivity with serum proteins from other species and were conjugated with either fluorescein (FITC) or lissamine rhodamine (LRh). Purified human CK-BBwas from Scripps, Polyvinol alcohol (Airvol 205) was from Air Products and Chemicals. Other reagents used were of the highest quality available from commercial sources.
Astrocyte cultures Astrocytes were purified from mixed glial cell cultures obtained from neonatal rat cerebral cortical hemispheres as described previously (Manos et al., '91). The cells were cultured in 90% Dulbecco's modified EagleiHams F-12 (DIF) medium (1:l)containing 10% calf serum, 15 mM
Abbreviations BSA CK CLSM FITC GFAP LRh PBS PI TBS
bovine serum albumin creatine kinase confocal laser-scanning microscopy fluorescein glial fibrillary acidic protein lissamine rhodamine phosphate-buffered saline propidium iodide tris-buffered saline
Hepes, and 1.2 g/l NaHC03. The cells were cultured for 10 days as a mixed glial population before the small, phasedark upper layer of process bearing cells were removed by the shear forces of a rotary shaker from the bed layer of astrocytes (McCarthy and de Vellis, '81).Use of this method and a medium supplemented with fetal bovine serum resulted in over 98% astrocytes in the cell population. Cultures of purified astrocytes were maintained in 98% DIF medium supplemented with 2% calf serum for 12 to 24 days until preparation for immunocytochemistry.
Immunocytochemistry and fluorescent staining Astrocytes were dislodged by treatment with trypsin and reseeded at a density of 1 x lo5 cells/cm2 onto glass coverslips in microwell plates. The cells were cultured for 24 hours in 90% D/F medium supplemented with 10% calf serum before processing for immunocytochemistry. The cells were fixed in 4% paraformaldehyde in PBS, pH 7.4 for 20 minutes at room temperature. The cells were permeabilized with 0.1% triton X-100 for 15 minutes and were blocked with 5% BSA, 0.1% NaNs in TBS, pH 7.5 for not less than 1.5 hours. The cells were incubated sequentially with antibodies for 60 to 90 minutes at RT. Antibodies were present at a dilution of 11200 unless otherwise indicated. Antitubulin serum was used at a dilution of l / l O O . All dilutions were prepared in 5% BSA, 0.1% NaN3 in TBS, pH 7.5. The fluorescent dye bisbenzimide 33258 (Hoechst stain) was added at a concentration 5 pgiml in the final antibody incubation solution. For confocal microscopy, some of the samples were treated with 1mgiml RNase A for 18 minutes at 37°C (Gray and Coffine, '79) to remove RNA before labeling with antibodies. Propidium iodide (PI) was used at a dilution of 11750 (of a 0.5 mgiml stock) except in RNase-treated samples, in which the dilution used was 11250 in the final antibody incubation solution (Jones and Kniss, '87). The coverslips were mounted onto slides with the aid of a solidifying mountant made from 17% Vinol205 polyvinol alcohol in TBS, pH 8.0, containing 6% 1,4diazabicyclo[2.2.2]octane (DABCO) in glycerol as an antifade compound. The anti-CK (rabbit) serum was treated with the purified human CK-BB protein to remove antibodies that react with the CK protein. We had previously determined that a 11400 to 11600 dilution of the anti-CK serum could inactivate 3 to 5 units of CK enzyme activity (unpublished data). This observation allowed us to determine the ratio of purified CK protein to anti-CK serum necessary to immunoadsorb the CK antibodies and significantly decrease the pattern of CK staining. A 11400 dilution of the anti-CK antisera in binding buffer (25 mM imidazole-Cl, pH 7.1, 75 mM NaCl, 2.5 mM dithiothreitol) was treated with 5 units (7 pg) of human CK protein for 5 minutes at room temperature, and 60 minutes at 7°C on a rocker-type shaker. The antiserum was centrifuged at 12,000 x g and the supernatant was then further treated as described above with 5 additional units of the purified human CK protein that had been denatured by boiling for 5 minutes with 0.1% sodium dodecyl sulphate. The mixture was centrifuged at 12,000 x g for 15 minutes and the supernatant was diluted 11500 in the BSA blocking buffer before use in immunocytochemistry. Control antiserum was prepared by substituting the CK buffer (400 mM NaC1,O.l mM EDTA, 10 mM P-mercaptoethanol) for the purified CK protein and was treated in an
CREATINE KINASE IN NUCLEAR ENERGY METABOLISM identical fashion as the antiserum treated with the CK protein. Astrocytes were incubated with either the CKtreated or buffer-treated antiserum for 4 hours at room temperature and labeled with the secondary antibody, rabbit anti-IgG-LRh, for 60 minutes at room temperature.
275
(Fig. 1E) and was minimal with antigoat IgG-LRh (Fig. 1F). Figure 1G is a phase contrast micrograph of the same field shown in Figure 1E,F. To address the possibility that the results were unique to the goat-derived anti-CK serum, the specificity of the CK staining pattern was examined by using an anti-CK serum Flu-orescencemicroscopy raised in rabbit. The rabbit anti-CK produced a similar For conventional fluorescence microscopy, photomicro- staining pattern as that observed from the CK antibody graphs were taken with a Nikon Microphot-FX microscope raised in goat; the cytoplasm appeared evenly stained with equipped with an episcodic fluorescence attachment. Fluo- the nuclear region of the cell staining intensely (Fig. 2A). rescent staining was observed through either a Fluor 20 x The cytoplasmic CK staining pattern was distinct from that (0.75 NA) or a Fluor 4 0 (0.85 ~ NA) objective and the seen with the cytoskeletal protein, tubulin, which produced following barrier filters: BA 420 for UV, BA 520-560 for a filamentous pattern (Fig. 2B). CK staining in the nuclear FITC and BA 590 for rhodamine. Photographs were taken region co-localized with the DNA-binding Hoechst stain (Fig. 2C) and can be compared to the position of the nuclei with T-MAX film, M A 400. Optical sections were obtained on a confocal laser- in the phase contrast micrograph (Fig. 2D). Nonspecific scanning microscope equipped with an argon ion laser binding of the secondary antibodies was barely detectable (model MRC-600, Bio-Rad Laboratories). Dual wavelength with antirabbit IgG-FITC (Fig. 2E) and was not detectable images were recorded simultaneously with the two channel with antimouse IgG-LRh (Fig. 2F). The same field is recording option, which uses the 514 nm laser line for presented in the phase contrast micrograph in Figure 2G. The specificity of the immunofluorescent labeling pattern excitation of both fluorochromes. Two independent transmission filters were used; one at 540 nm (for FITC) and 1 for CK was further examined by treating the CK antiserum long pass at 600 nm (for PI). The objective used in these (rabbit) with the purified CK protein before use of the studies was a Plan Apo 60x (1.4 NA). The confocal antiserum in immunocytochemistry. This treatment should apertures were set at 20% (for FITC) and 13% (for PI) remove antibodies from the antiserum that bind to the of the maximum aperture size possible. The interference purified CK protein. Figure 3 illustrates the immunofluoresdue to crossover of the fluorescent signal from one channel cent staining pattern produced by using either CK-treated to the other was determined in cells that were labeled with or buffer-treated antisera. To identify and locate the cell either anti-CK-IgG-FITC alone or PI alone. The crossover nucleus, the cells were also stained with the Hoechst stain of the PI signal to the FITC channel was found to be (Fig. 3A,D). The morphology and position of cells in this negligible. The crossover from the FITC channel to the PI region can be also observed in the phase contrast microchannel was small and was corrected for by subtracting the graphs (Fig. 3C,F). The control, buffer-treated, anti-CK maximum intensity of the crossover signal from the PI serum (Fig. 3B) produced a similar staining pattern for CK image. The rate of photobleaching of the sample during the as that observed in Figures 1A and 2A. Treatment of the collection of the images was minimal. Photographs were antiserum with 10 units of purified CK protein (14 p,g) taken with a Polaroid video image recorder on Kodak significantly reduced the intensity of the overall staining technical pan film (#2415). pattern and virtually eliminated the intensely staining nuclear region (Fig. 3E). It was evident from the results obtained from standard RESULTS fluorescence microscopy that antibodies against CK stained The subcellular localization of creatine kinase (CK) was the nuclear region of the cell. However, fine structural examined in cultured astrocytes by immunocytochemical details related to the specific location of CK within the approaches. The cells were stained with an anti-CK sera nucleus, or its association with the nuclear membrane (either a polyclonal antiserum raised in goat or in rabbit) could not be clearly discerned. This may be due to the and the staining pattern was compared with the staining background haze, which is created from fluorescent signals pattern for a cytoskeletal protein and a marker for the cell that arise from structures above and below the plane of nucleus. With conventional fluorescence microscopy, the focus and contribute to the final image. Confocal laserpattern for CK staining (with the goat antiserum) appeared scanning microscopy (CLSM) improves the image resoluuniform throughout the cell cytoplasm and was intense in tion in some samples over that obtained with the standard what appeared to be the nuclear region of the cell (Fig. 1A). fluorescence microscope; the point illumination-detection All cells present in the cultures stained positive for CK. The arrangement of the CLSM does not allow out-of-focus cytoplasmic CK staining pattern did not co-localizewith the fluorescence to contribute to the final image (White et al., staining pattern obtained with the cytoskeletal protein glial '87; Shuman et al., '89; Pawley, '90; Wilson, '90). In fibrillary acidic protein (GFAP, Fig. 1B). GFAP is an addition, the use of a confocal microscope allows us to intermediate filament protein that is specific for astrocytes exclude the possibility that the nuclear localization of CK (Bignami et al., '72; Debus et al., '83). Purified astrocyte was an artifact due to the greater thickness of the cell in the populations cultured in medium supplemented with calf region of the nucleus. Examples of confocal images are serum were routinely over 80% GFAP positive. The CK presented in Figure 4. As a control for the specificity of the positive nuclear region co-localized with the staining pat- primary antiserum, the cells were labeled with nonimmune tern obtained from the Hoechst stain, a DNA-binding rabbit serum instead of the anti-CK serum derived from ligand that binds adenine-thymine regions with high afin- rabbit. Replacement of anti-CK by nonimmune rabbit ity (Muller and Gautier, '75). Cell morphology and position serum resulted in weak staining (Fig. 4A). Three optical can also be observed in a phase contrast micrograph of the sections were taken sequentially through a fixed monolayer same field (Fig. 1D). Nonspecific binding of the secondary of astrocytes that had been double-labeled with anti-CK antibodies was not detectable with antimouse IgG-FITC and propidium iodide (PI).P I stains total nucleic acid and is
-
-
P. MANOS AND J. EDMOND
276
Fig. 1. Fluorescent localization of creatine kinase (CK), glial fibrillary acidic protein (GFAP)and nuclei in cultured astrocytes. Astrocytes were treated sequentially with the following antibodies (as indicated in Materials and Methods): goat anti-CK, donkey antigoat I&-LRh (A), mouse anti-GFAF', donkey antimouse I&-FITC (B).The final antibody solution contained the Hoechst nuclear stain (C). Corresponding field as seen with phase contrast optics (D).CK has a cytoplasmic staining
-
specific for DNA in RNase-treated samples. Each section represents a depth of focus of 0.7 pm (Wells et al., '891, and the images were recorded simultaneously in the same sections. CK staining is present in the cell nucleus in each optical section taken through the monolayer (Fig. 4B,1,2,3, left panel). CK positive staining in the nucleus appeared in
pattern that is distinct from GFAP and a nuclear staining pattern that co-localizes with the Hoechst stain. Photomicrographs were obtained from a conventional fluorescence microscope. To determine the extent of nonspecific binding of the secondary antibodies, control wells were incubated with 5%donkey serum instead of the primary antibodies and treated sequentially with antimouse I&-FITC (El and antigoat IgGLRh (F).Corresponding phase contrast field (GI. Bar : 25 pm.
the same sections through the nucleus as PI-labeled DNA (Fig. 4B 1,2,3,right panel), which demonstrates that CK is located within the nucleus. However, the labeling pattern of anti-CK and PI were clearly different in each nuclear section. Under conditions containing RNase, PI stains double-stranded DNA whereas CK staining has a particu-
CREATINE KINASE IN NUCLEAR ENERGY METABOLISM
Fig. 2. Fluorescent localization of creatine kinase (CK), tubulin and nuclei in cultured astrocytes. Astrocytes were sequentially incubated with the following antibodies (as indicated in Materials and Methods): rabbit anti-CK, donkey antirabbit I&-FITC (A), mouse antitubulin, donkey antimouse IgGLRh (B).The final antibody solution contained the Hoechst stain (C). Corresponding field as seen with phase contrast optics (D).These anti-CK antibodies, raised in rabbit, produce a similar
277
labeling pattern as that seen with the goat anti-CK antibodies (Fig. 1). Photomicrographs were obtained from a conventional fluorescence microscope. To determine the extent of nonspecific binding of the secondary antibodies, control wells were incubated with 5% donkey serum instead of the primary antibodies and treated with the secondary antibodies: antirabbit IgG-FITC (El and antimouse IgG-LRh (F). Corresponding phase contrast field ( G ) .Bar = 25 km.
278
P. MANOS AND J. EDMOND
Fig. 3. Removal of the pattern of CK staining by preadsorption of anti-CK serum with purified CK protein. Astrocytes were incubated with anti-CK serum (rabbit) that had been treated with either a CK-buffer (B) or with the purified CK protein (El as described in Materials and Methods. The cells were then incubated with antirabbit IgG-LRh containing the Hoechst stain. (A,D) Hoechst nuclear stain;
(C) is the correspondingphase contrast field as in (A) and (B); (F)is the corresponding phase contrast field as in (D) and (El.(A,B,C) were photographed, developed, and printed under identical conditions (film speed settings and exposure times) as (D,E,F). Photomicrographs were obtained from a standard fluorescence microscope. Bar = 50 km.
late, granular-likepattern throughout the nucleus, which is distinct from the pattern of staining for DNA. The pattern of CK staining was also compared with the staining pattern produced by labeling total nucleic acid (Fig. 5 ) . Dual channel images were recorded in a similar manner as described for Figure 4, except the cultures were not pretreated with RNase. PI therefore binds both doublestranded DNA and RNA. Six optical sections were taken through the cell monolayer. The first section was taken with the plane of focus just above the cell monolayer (Fig. 5A). The arrows indicate a representative astrocyte in which the CK stained nucleus (left panel) and PI-stained nucleic acids (right panel) can be very clearly seen (Fig. 5A,B) as the nucleus protrudes from the plane of the cell monolayer. P I staining is particularly intense in the subnuclear RNA-containing organelle, the nucleolus (Fig. 5C). Many of the nuclei contain more than one nucleolus. As successively deeper sections are taken through the monolayer, the amount of cytoplasmic cellular detail increases with both CK and PI staining (Fig. 5D,E). PI produces a distinct, filamentous pattern of staining. The nuclear CK and PI staining pattern in the cell indicated by the arrow becomes less apparent as sections are taken below the plane of focus that contains the nucleus (Fig. 5E,F). The pattern of CK staining in each section through the nucleus can be very clearly seen. CK is present throughout the nucleus and it does not stain the nucleolus. The particulate CK staining pattern is suggestive of a nucleoplasmic (karyoplasmic) distribution.
DISCUSSION Creatine kinase, an enzyme important in energy homeostasis, is distributed uniformly throughout the cytoplasm in cultured astrocytes and is especially intense in the nucleus. This finding of intense immunoreactive CK-B in the astrocyte nucleus was unexpected and, to our knowledge, has not been reported previously. The use of confocal microscopy has enabled us to elucidate the nuclear substructure of CK. CK has a punctate, granular-like pattern throughout the nucleus, which indicates that CK is a nucleoplasmic protein. CK is not found in the nucleolus and has a distinctly different staining pattern from the staining pattern produced by labeling either DNA or RNA. We believe the presence of intensely staining CK in the nucleus indicates that CK and the creatine-creatine phosphate energy pathway may provide, in part, the ATP necessary for the energy dependent processes of nuclear function. We used several immunological controls to assess the specificity of antibody binding. Nigg ('88) has suggested that irregular and nonspecific binding can be a potential problem, particularly with nuclear antigens, and can occur with a variety of immune or nonimmune sera. We found that two different antisera raised against the CK protein, in different species, produced a similar pattern of CK staining. The distribution of the CK label and the intense nuclear staining were nearly identical with both the goat and the rabbit antisera, even though different fluorochromes were conjugated to the second antibody (in Fig. lA, CK was
CREATINE KINASE IN NUCLEAR ENERGY METABOLISM
279
Fig. 4. Serial optical sections of immunostained CK and propidium iodide-stained DNA,using confocal microscopy. Control experiments in which astrocytes were treated with nonimmune rabbit serum instead of anti-CK and labeled with antirabbit IgG-FITC resulted in weak staining (A, left panel). Crossover of the FITC signal into the PI channel was not detectable. Bar = 25 bm (A,right panel). B.Astrocytes were treated with RNase before labeling with anti-CK serum (rabbit) and antirabbit
IgG-FITC (left panel). The second antibody solution contained PI (right panel). See Materials and Methods for more detail. B-1,2,3 represent three consecutive optical sections taken with a depth of focus of -0.7 pm and a distance of 0.5 ym along the Z axis. CK has a nucleoplasmic distribution that is distinct from DNA. Five scans were accumulated and averaged for each image. Bar = 10 pm.
detected with LRh; in Fig. 2A, CK was detected with FITC). In addition, the same pattern of CK staining was observed for the rabbit derived anti-CK with both FITC (Figs. 2A, 5) and LRh (Fig. 3B) as the fluorochrome. These controls eliminated the possibility of artifacts due to differences in binding of the secondary antibodies or intensity differences in the fluorochromes. Nonspecific binding of the fluoro-
chrome-conjugated secondary antibodies was also minimal (Figs. lE,F,G, 2E,F,G). We were also able to substantially decrease CK staining in the cytoplasm and nucleus by pre-treating the antiserum with the purified CK protein (Fig. 3). Nonspecific labeling with nonimmune rabbit serum was also minimal (Fig. 4A). These data argue convincingly that the labeling pattern produced with anti-CK is
280
P. MANOS AND J. EDMOND
-
Fig. 5. Serial optical sections of immunostained CK and propidium iodide stained total nucleic acid, using confocal microscopy. Astrocytes were labeled with anti-CK (rabbit) serum and antirabbit IgG-FITC (left panel) as described in Material and Methods. The second antibody solution contained PI (right panel). (A-F) Six consecutive optical
sections were taken with a depth of focus of 0.7 Fm and a distance of 1.0 km along the Z axis. The pattern of CK stainingis distinct from the pattern of staining for either RNA or DNA. Note the intense RNA stained nucleoli (right panels). Five scans were accumulated and averaged for each image. Bar = 25 km.
specific for CK and that CK is present in the nucleus of astrocytes cultured under these conditions. Several groups have examined the immunohistochemical localization of CK in brain sections (Yoshimine et al., '83; Kato et al., '86; Ikeda and Tomonaga, '87). CK immunoreactivity was found in the astrocyte cytoplasm but was not reported to stain the nucleus, nor was nuclear staining obvious from the photomicrographs presented. The lack of detectable CK in the astrocyte nucleus in the above studies may be attributable to the use of organic fixatives like methanol and ethanol. These fixatives have been reported
to extract or destroy nuclear antigens (Nigg, '88). In our experience, the use of methanol as a fixative greatly diminishes the intensity of total CK staining. Alternatively, the lack of detectable CK in the nucleus of astrocytes in brain sections may result from the fact that the sections were taken from adult brains, whereas the cultured astrocytes were obtained from neonatal brains. The localization of CK in the nucleus may result from the particular energy requirements of the nucleus during a specific period in development or in certain physiological states. In this regard, astrocytes in this study were reseeded 24 hours
CREATINE KINASE IN NUCLEAR ENERGY METABOLISM before preparation for immunocytochemistry. The replating of the cells may have an effect on the expression and/or subcellular localization of CK as the cells respond to the trauma of subculturing. Many enzymes within the nucleus require high energy phosphate in the form of (d)NTPs or, specifically,ATP. The (d)NTPs serve as substrates for both RNA and DNA directed polymerases in the synthesis of nucleic acids, and they are believed to provide energy for macromolecular transport (Newport and Forbes, '87). ATP is required, specifically, by Poly (A) polymerase and other nuclear enzymes such as ligases, topoisomerases, and helicases. (d)NTP's are generated by the addition of a phosphate to the corresponding NDP by cytoplasmic NDP kinases, using ATP as the phosphate donor. Cytoplasmic ATP can be generated from substrate level phosphorylation (anaerobic) or oxidative phosphorylation (aerobic). The nucleus, as an organelle, does not contain any primary energy generating mechanisms and therefore depends on the recruitment of ATP and other high energy phosphate compounds from the cytoplasm. Creatine phosphate, as an energy donor, may provide a local and sustained concentration of ATP at specific sites in the nucleus. Cellular ATPases are unable to use the high energy phosphate in creatine phosphate unless CK is present. Nuclear ATPases, therefore, need not compete with other cellular ATPases for ATP, as only ATPases associated with CK are able to use the high energy phosphate in creatine phosphate. Thus, CK and the creatinecreatine phosphate energy "shuttle" may form a link between the generation of energy by oxidative metabolism in the cytoplasm with substrate level phosphorylation in the astrocyte nucleus. Macromolecules are believed to be transported across the nuclear envelope via the nuclear pore by an active transport mechanism (Newport and Forbes, '86). The nuclear pore complex within the nuclear envelope is thought to be comprised of the machinery for the translocation of macromolecules, the contractile machinery, and the ATP hydrolyzing machinery. The export of messenger ribonucleoprotein particles (Agutter, '88) and ribosomes (Schumm et al., '79) and import of some nuclear proteins (Newmeyer, '90; Wagner et al., '90) has been shown to be energy dependent. ATPase activity has been detected histochemically on the nuclear envelope (Sikstrom et al., '76; Vorbrodt and Maul, '80; Fox et al., 'Bl), in discrete areas within the nucleolus (Fox et al., '81; Sikstrom et al., '761, with structures known to contain ribonucleoproteins (Vorbrodt and Maul, '801, and in deposits scattered throughout the nucleoplasm (Sikstrom et al., '76). In addition, ATPases are believed to be associated with nuclear contractile proteins (LeStourgeon, '78; Berrios and Fisher, '86). Thus nuclear CK may provide high energy phosphate in the region of these ATPases that are involved in macromolecular transport, contractile events, or other movements within the nucleus. The use of confocal microscopy as a detection method in these immunofluorescent studies has allowed us to elucidate structural details of CK in the astrocyte nucleus in greater detail than possible with conventional fluorescence microscopy. CK appears to have a ''soluble'' nucleoplasmic distribution. The pattern of CK staining in the nucleus was intriguing, and nonhomogeneous with the pattern of staining of either DNA or RNA. In this regard, a similar nucleoplasmic fluorescent staining pattern, which was incongruent with the staining pattern for DNA, has been observed for the product of the fos protooncogene (Nishiku-
281
ra and Murray, '87; Shuman et al., '89). We are very interested in the role CK serves in the astrocyte nucleus and in determining whether CK binds to DNA or interacts with other nuclear macromolecules.
ACKNOWLEDGMENTS We thank Justine Tseng for technical assistance during the preliminary stages of this work, Dr. Guy K. Bryan for critical review of the manuscript, and Drs. Michael S. Levine and Dorwin Birt for their assistance with the confocal microscope and computer imaging system. This work was supported by HD 06576.
LITERATURE CITED Agutter, P.S. (1988) Nucleo-cytoplasmic transport of mRNA Its relationship to RNA metabolism, subcellular structures and other nucleoplasmic exchanges. Prog. Mol. Subcell. Biol. 10:15-96. Austead, N., R.A. Korsak, J.W. Morrow, and J. Edmond (1991) Fatty acid oxidation and ketogenesis by astrorvtes in primary culture. J. Neurochem. 56:1376-1386. Berrios, M., and P.A. Fisher (1986) A myosin heavy chain-like polypeptide is associated with the nuclear envelope in higher eukaryotic cells. J. Cell Biol. 103:711-724. Bessman, S.P., and P.J. Geiger (1981) Transport of energy in muscle: The phosphorylcreatine shuttle. Science 211:448-452. Bessman, S.P., and C.L. Carpenter (1985) The creatine-creatine phosphate energy shuttle. Ann. Rev. Biochem. 54.832462. Bignami, A., L.F. Eng, A. Dahl, and C.T. Uyeda (1972) Localization of the glial fibrigary acidic protein in astrocytes by immunofluorescence. Brain Res. 43:429435. Debus, E., K. Weber, and M. Osborn (1983) Monoclonal antibodies specific for glial fibrillary acidic (GFA) protein and for each of the neurofilament triplet polypeptides. Differentiation 25: 193-203. Edmond, J., R.A. Robbins, J.D. Bergstrom, R.A. Cole, and J. deVellis (1987) Capacity for substrate utilization in oxidative metabolism by neurons, astrocytes and oligodendrocytes from developing brain in primary culture. J. Neurosci. Res. 18:551-561. Fox, N., C. Fernandez, and G.P. Studzinski (1981) Visualization of nucleolar substructure in cultured human fibroblasts by magnesium-activated adenosine triphosphate reaction. J. Histochem. Cytochem. 29: 11151120. Gray, J.W., and P. Coffin0 (1979) Cell cycle analysis by flow cytometry. 'Methods Enzymol58:233-248. Hawkins, R. (1985) Cerebral Energy Metabolism. In D.W. McCandless (ed): Cerebral Energy Metabolism and Metabolic Encephalopathy. New York: Plenum Press, pp. 3-23. Hertz, L. (1981) Features of astrocytic function apparently involved in the response of central nervous tissue to ischemia-hypoxia. J. Cereb. Blood Flow Metab. 1:143-153. Ikeda, K., and M. Tomonaga (1987) The presence of creatine kinase (CK)-immunoreactive neurons in the zona incerta and lateral hypothalamic areas of the mouse brain. Brain Res. 435348-350. Jones, K.H., and D.A. Kniss (1987) Propidium iodide as a nuclear counterstain for immunofluorescence studies on cells in culture. J. Histochem. Cytochem. 35:123-125. Kato, K., F. Suzuki, A. Shimizu, H. Shinohara, and R. Semba (1986) Highly sensitive immunoassay for rat brain-type creatine kinase: Determination in isolated Purkinje cells. J. Neurochem. 46:1783-1786. Kenyon, G.L., and G.H. Reed (1983) Creatine kinase: Structure-activity relationships. Adv. Enzymol. 54:367-426. LeStourgeon, W.M. (1978) The occurrence of contractile proteins in nuclei and their possible function. In H. Busch (ed):The Cell Nucleus. Vol. VI. New York: Academic Press, Inc, pp 305-326. Manos, P., G.K. Bryan, and J. Edmond (1991) Creatine kinase in postnatal rat brain development and in cultured neurons, astrocytes and oligodendrocytes. J. Neurochem. 56:2101-2107. Martinez-Hernandez, A,, K.P. Bell, and M.D. Norenberg (1977) Glutamine synthase: Glial localization in brain. Science 195:1356-1358. McCarthy, K.D., and J. de Vellis (1980) Preparation of separate astroglia and oligodendroglia cell cultures from rat cerebral tissue. J. Cell Biol. 85:890-902.
282 Meyer, R.A., H.L. Sweeney, and M.J. Kushmerick (1984) A simple analysis of the “phosphocreatine shuttle.” Am. J. Physiol. 15:C365-C389. Muller, W., and F. Gautier (1975) Interactions of heteroaromatic compounds with nucleic acids. A.T-specific non-intercalating DNA ligands. Eur. J. Biochem. 54:385-394. Newmeyer, D.D. (1990) Nuclear import in vitro. Prog. Mol. Subcell. Biol. IIr12-50. Newport, J.W., and D.J. Forbes (1987)The nucleus: Structure, function and dynamics. Ann. Rev. Biochem. 56t535-565. Nigg, E.A. (1988)Nuclear function and organization: The potential of immunochemical approaches. Int. Rev. Cytol. I IOr27-92. Nishikura, K., and J.M. Murray (1987) Antisense RNA of proto-oncogene c-fos blocks renewed growth of quiescent 3T3 cells. Mol. Cell Biol. 73339-649. Pawley, J.B. (1990)Handbook of Biological Confocal Microscopy. New York: Plenum Press, 232 pp. Schumm, D.E., M.A. Niemann, T. Palayoor, and T.E. Webb (1979) In vivo equivalence of a cell-free system from rat liver for ribosomal RNA processing and transport. J. Biol. Chem. 254r12126-12130. Shuman, H., J.M. Murray, and C. DiLullo (1989) Confocal microscopy: An overview. Biotechniques 7: 154-163. Sikstrom, R., J. Lanoix, and J.J.M. Bergeron (1976) An enzymatic analysis of a nuclear envelope fraction. Biochem. Biophys. Acta 448:88-102. Tombes, R.M., and B.M. Shapiro (1985) Metabolitechanneling: aphosphorylcreatine shuttle to mediate high energy phosphate transport between sperm mitochondrion and tail. Cell 42r325-334.
P. MANOS AND J. EDMOND Vorbrodt, A,, and G.G. Maul (1980) Cytochemical studies on the relation of nucleoside triphosphatase activity to ribonucleoproteins in isolated rat liver nuclei. J. Histochem. Cytochem. 28t27-35. Wagner, P., J. Kunz, A. Koller, and M.N. Hall (1990) Active transport into the nucleus. FEBS Letters 275:l-5. Wallimann, T., and H. Eppenberger (1990) The subcellular compartmentation of creatine kinase isoenzymes as a precondition for a proposed phosphoryl-creatine circuit. Prog. Clin. Biol. Res. 344:877-889. Wallimann, T., M.Wyss, D. Brdiczka, K. Nicolay, and H. Eppenberger (1992) Intracellular compartmentation, structure and function of creatine kinase isoenzymes in tissues with high and fluctuating energy demands: The “phosphocreatine circuit” for cellular energy homeostasis. Biochem. J. 281t21-40. Watts, D.C. (1973) Creatine kinase (adenosine 5’-triphosphate-~reatine phosphotransferase). The Enzymes 8:383-455. Wells, K.S., D.R. Sandison, J. Stricker, and W.W. Webb (1990) Quantitative fluorescence imaging with laser scanning confocal microscopy. In J. Pawley (ed): The Handbook of Biological Confocal Microscopy. New York: Plenum Press, pp. 27-39. White, J.G., W.B. Amos, and M. Fordham (1987) An evaluation of confocal versus conventional imaging of biological structures by fluorescence light microscopy. J. Cell Biol. 105t41-48. Wilson, T. (1990) Confocal Microscopy. New York: Academic Press, 426 pp. Yoshimine, T., K. Morimoto, H.A. Homburger, and T. Yanagihara (1983) Immunochemical localization of creatine kinase-BB isoenzyme in human brain: Comparison with tubulin and astroprotein. Brain Res. 265r101108.
Immunofluorescent Analysis of Creatine Kinase in Cultured Astrocytes by Conventional and Confocal Microscopy: A Nuclear Localization PATRICIA MANOS AND JOHN EDMOND Department of Biological Chemistry and Mental Retardation Research Center, UCLA School of Medicine, Los Angeles, California 90024
ABSTRACT The subcellular localization of creatine kinase (CK) was examined in primary cultures of astrocytes with immunofluorescent labeling methods and detection by both standard fluorescence microscopy and confocal laser-scanning microscopy. With conventional microscopy, the pattern of CK staining was uniform throughout the cell cytoplasm and appeared to stain the nuclear region intensely. Staining of CK in the nuclear region co-localized with the DNAspecific Hoechst nuclear stain. CK produced a diffuse cytoplasmic staining pattern that was different from the staining pattern produced by the cytoskeletal proteins glial fibrillary acidic protein and tubulin, both of which showed a filamentous cytoskeletal network that excluded the nucleus. To examine the structural details of CK in the nuclear region, serial optical sections were taken through the cell monolayer with a confocal microscope. The cells were immunostained for CK, and the CK-staining pattern was compared with the staining pattern produced by propidium iodide, which is specific for DNA in RNase-treated samples and stains total nucleic acid in untreated samples. CK staining was present within the nucleus in each section taken through the monolayer. The nucleolus did not stain for CK. The pattern of CK staining in the nucleus (and cytoplasm) was distinctly different from the staining pattern of either DNA or total nucleic acid. Nuclear CK appeared to have a granular, particulate pattern, which is suggestive of a nucleoplasmic distribution. o 1992 Wiley-Liss, Inc. Key words: creatine phosphokinase, cerebral energy metabolism, glia
The mammalian central nervous system has a high rate of oxidative metabolism and is believed to use, a large portion of the total body energy requirement (Hawkins, ’85).Creatine, creatine phosphate, and creatine kinase (CK, EC 2.7.3.2)play an important role in brain energy homeostasis, in part by maintaining cellular ATP levels. CK catalyses the reversible transfer of a high energy phosphate between ATP and creatine phosphate and is high in tissues with high energy requirements like brain, skeletal muscle, and heart. CK is a dimeric enzyme with four different molecular forms (for reviews, see: Watts, ’73; Kenyon and Reed, ’83; Bessman and Carpenter, ’85; Wallimann et al., ’92). One isoenzyme is located on the inner mitochondrial membrane and functions in respiratory control. The other three isoforms are believed to be primarily cytoplasmic and are expressed in a tissue-specific fashion. The “B” subunit type is the predominant form in brain, whereas the “M” subunit type is the predominant form in skeletal muscle; the heterodimer MB form is characteristic for heart. Cytoplasmic CK is believed to provide a sustained, local, and
o 1992 WILEY-LISS, INC.
high concentration of ATP at cellular sites of high energy utilization. The CK reaction has been described as a near-equilibrium reaction that functions as an energy reservoir, or as a buffer for ATP levels (Watts, ’73; Kenyon and Reed, ’83; Meyer et al., ’84). Alternatively, arguments based on the metabolic compartmentation of adenine nucleotides and the CK isoenzymes have led to the idea of an energy “shuttle” (Bessman and Geiger, ’81;Wallimann and Eppenberger, ’90). High energy phosphate in the form of creatine phosphate is targeted to specific sites in the cell where cytoplasmic CK is associated with an ATPase. Cellular ATPases are unable to use the high energy phosphate in creatine phosphate unless CK is present. Creatine is then rephosphorylated by the mitochondrial isoenzyme. The creatine-creatine phosphate energy shuttle has been well Accepted August 14,1992 Address reprint requests to Dr. Patricia Manos, Department of Biological Chemistry, 33-257 Center for the Health Sciences, UCLA School of Medicine, Los Angeles, CA 90024.
P. MANOS AND J. EDMOND
274 described in muscle (Bessman and Geiger, '81; Meyer et al., '84; Bessman and Carpenter, '85; Wallimann and Eppenberger, '90; Wallimann et al., '92) and sperm (Tombes and Shapiro, '85). CK activity increases rapidly in postnatal brain development and is present in homogeneous populations of cultured neurons, oligodendrocytes, and astrocytes (Manos et al., '91). The metabolic activity of astrocytes is high and is thought to be comparable to that of neurons (Hertz, '81). Many metabolic pathways, especially those related to energy metabolism, are unique to astrocytes. Astrocytes function in glutamate uptake and ammonia fixation (Martinez-Hernandez et al., '77; Hertz, '81). Astrocytes have the capacity to oxidize fatty acids (Edmond et al., '87) as well as synthesize ketone bodies (Austead et al., '91) in support of the energy requirements in developing brain. Because of the importance of astrocytes in cerebral energy metabolism, we were interested in examining the subcellular localization of CK in cultured astrocytes. With conventional fluorescent microscopy, immunoreactive CK was found to be dispersed throughout the cytoplasm and to stain the nuclear region of the cell intensely. The staining pattern of CK was distinct from the staining pattern of cytoskeletal proteins and co-localized with the Hoechst stain. Serial optical sections taken through the cell monolayer with a confocal microscope indicate that CK is present uniformly throughout the nucleus, does not co-localize with either DNA or RNA, and has a punctate, or granular, pattern that indicates it is located in the nucleoplasm.
MATERIAL AND METHODS Rats were purchased from Bantam and Kingman; cell culture medium was from GIBCO; Calf Serum (Lot #2 151682) was from Hyclone Laboratories. The antibodies were obtained commercially from the following sources: rabbit antihuman CK-BB (Calbiochem); goat antihuman CK-BB (Chemicon); mouse anti-GFAP (Boheringer Mannheim); mouse antitubulin (Boheringer Mannheim). The secondary antibodies, purchased from Jackson Laboratories, were affinity purified IgG with minimal cross reactivity with serum proteins from other species and were conjugated with either fluorescein (FITC) or lissamine rhodamine (LRh). Purified human CK-BBwas from Scripps, Polyvinol alcohol (Airvol 205) was from Air Products and Chemicals. Other reagents used were of the highest quality available from commercial sources.
Astrocyte cultures Astrocytes were purified from mixed glial cell cultures obtained from neonatal rat cerebral cortical hemispheres as described previously (Manos et al., '91). The cells were cultured in 90% Dulbecco's modified EagleiHams F-12 (DIF) medium (1:l)containing 10% calf serum, 15 mM
Abbreviations BSA CK CLSM FITC GFAP LRh PBS PI TBS
bovine serum albumin creatine kinase confocal laser-scanning microscopy fluorescein glial fibrillary acidic protein lissamine rhodamine phosphate-buffered saline propidium iodide tris-buffered saline
Hepes, and 1.2 g/l NaHC03. The cells were cultured for 10 days as a mixed glial population before the small, phasedark upper layer of process bearing cells were removed by the shear forces of a rotary shaker from the bed layer of astrocytes (McCarthy and de Vellis, '81).Use of this method and a medium supplemented with fetal bovine serum resulted in over 98% astrocytes in the cell population. Cultures of purified astrocytes were maintained in 98% DIF medium supplemented with 2% calf serum for 12 to 24 days until preparation for immunocytochemistry.
Immunocytochemistry and fluorescent staining Astrocytes were dislodged by treatment with trypsin and reseeded at a density of 1 x lo5 cells/cm2 onto glass coverslips in microwell plates. The cells were cultured for 24 hours in 90% D/F medium supplemented with 10% calf serum before processing for immunocytochemistry. The cells were fixed in 4% paraformaldehyde in PBS, pH 7.4 for 20 minutes at room temperature. The cells were permeabilized with 0.1% triton X-100 for 15 minutes and were blocked with 5% BSA, 0.1% NaNs in TBS, pH 7.5 for not less than 1.5 hours. The cells were incubated sequentially with antibodies for 60 to 90 minutes at RT. Antibodies were present at a dilution of 11200 unless otherwise indicated. Antitubulin serum was used at a dilution of l / l O O . All dilutions were prepared in 5% BSA, 0.1% NaN3 in TBS, pH 7.5. The fluorescent dye bisbenzimide 33258 (Hoechst stain) was added at a concentration 5 pgiml in the final antibody incubation solution. For confocal microscopy, some of the samples were treated with 1mgiml RNase A for 18 minutes at 37°C (Gray and Coffine, '79) to remove RNA before labeling with antibodies. Propidium iodide (PI) was used at a dilution of 11750 (of a 0.5 mgiml stock) except in RNase-treated samples, in which the dilution used was 11250 in the final antibody incubation solution (Jones and Kniss, '87). The coverslips were mounted onto slides with the aid of a solidifying mountant made from 17% Vinol205 polyvinol alcohol in TBS, pH 8.0, containing 6% 1,4diazabicyclo[2.2.2]octane (DABCO) in glycerol as an antifade compound. The anti-CK (rabbit) serum was treated with the purified human CK-BB protein to remove antibodies that react with the CK protein. We had previously determined that a 11400 to 11600 dilution of the anti-CK serum could inactivate 3 to 5 units of CK enzyme activity (unpublished data). This observation allowed us to determine the ratio of purified CK protein to anti-CK serum necessary to immunoadsorb the CK antibodies and significantly decrease the pattern of CK staining. A 11400 dilution of the anti-CK antisera in binding buffer (25 mM imidazole-Cl, pH 7.1, 75 mM NaCl, 2.5 mM dithiothreitol) was treated with 5 units (7 pg) of human CK protein for 5 minutes at room temperature, and 60 minutes at 7°C on a rocker-type shaker. The antiserum was centrifuged at 12,000 x g and the supernatant was then further treated as described above with 5 additional units of the purified human CK protein that had been denatured by boiling for 5 minutes with 0.1% sodium dodecyl sulphate. The mixture was centrifuged at 12,000 x g for 15 minutes and the supernatant was diluted 11500 in the BSA blocking buffer before use in immunocytochemistry. Control antiserum was prepared by substituting the CK buffer (400 mM NaC1,O.l mM EDTA, 10 mM P-mercaptoethanol) for the purified CK protein and was treated in an
CREATINE KINASE IN NUCLEAR ENERGY METABOLISM identical fashion as the antiserum treated with the CK protein. Astrocytes were incubated with either the CKtreated or buffer-treated antiserum for 4 hours at room temperature and labeled with the secondary antibody, rabbit anti-IgG-LRh, for 60 minutes at room temperature.
275
(Fig. 1E) and was minimal with antigoat IgG-LRh (Fig. 1F). Figure 1G is a phase contrast micrograph of the same field shown in Figure 1E,F. To address the possibility that the results were unique to the goat-derived anti-CK serum, the specificity of the CK staining pattern was examined by using an anti-CK serum Flu-orescencemicroscopy raised in rabbit. The rabbit anti-CK produced a similar For conventional fluorescence microscopy, photomicro- staining pattern as that observed from the CK antibody graphs were taken with a Nikon Microphot-FX microscope raised in goat; the cytoplasm appeared evenly stained with equipped with an episcodic fluorescence attachment. Fluo- the nuclear region of the cell staining intensely (Fig. 2A). rescent staining was observed through either a Fluor 20 x The cytoplasmic CK staining pattern was distinct from that (0.75 NA) or a Fluor 4 0 (0.85 ~ NA) objective and the seen with the cytoskeletal protein, tubulin, which produced following barrier filters: BA 420 for UV, BA 520-560 for a filamentous pattern (Fig. 2B). CK staining in the nuclear FITC and BA 590 for rhodamine. Photographs were taken region co-localized with the DNA-binding Hoechst stain (Fig. 2C) and can be compared to the position of the nuclei with T-MAX film, M A 400. Optical sections were obtained on a confocal laser- in the phase contrast micrograph (Fig. 2D). Nonspecific scanning microscope equipped with an argon ion laser binding of the secondary antibodies was barely detectable (model MRC-600, Bio-Rad Laboratories). Dual wavelength with antirabbit IgG-FITC (Fig. 2E) and was not detectable images were recorded simultaneously with the two channel with antimouse IgG-LRh (Fig. 2F). The same field is recording option, which uses the 514 nm laser line for presented in the phase contrast micrograph in Figure 2G. The specificity of the immunofluorescent labeling pattern excitation of both fluorochromes. Two independent transmission filters were used; one at 540 nm (for FITC) and 1 for CK was further examined by treating the CK antiserum long pass at 600 nm (for PI). The objective used in these (rabbit) with the purified CK protein before use of the studies was a Plan Apo 60x (1.4 NA). The confocal antiserum in immunocytochemistry. This treatment should apertures were set at 20% (for FITC) and 13% (for PI) remove antibodies from the antiserum that bind to the of the maximum aperture size possible. The interference purified CK protein. Figure 3 illustrates the immunofluoresdue to crossover of the fluorescent signal from one channel cent staining pattern produced by using either CK-treated to the other was determined in cells that were labeled with or buffer-treated antisera. To identify and locate the cell either anti-CK-IgG-FITC alone or PI alone. The crossover nucleus, the cells were also stained with the Hoechst stain of the PI signal to the FITC channel was found to be (Fig. 3A,D). The morphology and position of cells in this negligible. The crossover from the FITC channel to the PI region can be also observed in the phase contrast microchannel was small and was corrected for by subtracting the graphs (Fig. 3C,F). The control, buffer-treated, anti-CK maximum intensity of the crossover signal from the PI serum (Fig. 3B) produced a similar staining pattern for CK image. The rate of photobleaching of the sample during the as that observed in Figures 1A and 2A. Treatment of the collection of the images was minimal. Photographs were antiserum with 10 units of purified CK protein (14 p,g) taken with a Polaroid video image recorder on Kodak significantly reduced the intensity of the overall staining technical pan film (#2415). pattern and virtually eliminated the intensely staining nuclear region (Fig. 3E). It was evident from the results obtained from standard RESULTS fluorescence microscopy that antibodies against CK stained The subcellular localization of creatine kinase (CK) was the nuclear region of the cell. However, fine structural examined in cultured astrocytes by immunocytochemical details related to the specific location of CK within the approaches. The cells were stained with an anti-CK sera nucleus, or its association with the nuclear membrane (either a polyclonal antiserum raised in goat or in rabbit) could not be clearly discerned. This may be due to the and the staining pattern was compared with the staining background haze, which is created from fluorescent signals pattern for a cytoskeletal protein and a marker for the cell that arise from structures above and below the plane of nucleus. With conventional fluorescence microscopy, the focus and contribute to the final image. Confocal laserpattern for CK staining (with the goat antiserum) appeared scanning microscopy (CLSM) improves the image resoluuniform throughout the cell cytoplasm and was intense in tion in some samples over that obtained with the standard what appeared to be the nuclear region of the cell (Fig. 1A). fluorescence microscope; the point illumination-detection All cells present in the cultures stained positive for CK. The arrangement of the CLSM does not allow out-of-focus cytoplasmic CK staining pattern did not co-localizewith the fluorescence to contribute to the final image (White et al., staining pattern obtained with the cytoskeletal protein glial '87; Shuman et al., '89; Pawley, '90; Wilson, '90). In fibrillary acidic protein (GFAP, Fig. 1B). GFAP is an addition, the use of a confocal microscope allows us to intermediate filament protein that is specific for astrocytes exclude the possibility that the nuclear localization of CK (Bignami et al., '72; Debus et al., '83). Purified astrocyte was an artifact due to the greater thickness of the cell in the populations cultured in medium supplemented with calf region of the nucleus. Examples of confocal images are serum were routinely over 80% GFAP positive. The CK presented in Figure 4. As a control for the specificity of the positive nuclear region co-localized with the staining pat- primary antiserum, the cells were labeled with nonimmune tern obtained from the Hoechst stain, a DNA-binding rabbit serum instead of the anti-CK serum derived from ligand that binds adenine-thymine regions with high afin- rabbit. Replacement of anti-CK by nonimmune rabbit ity (Muller and Gautier, '75). Cell morphology and position serum resulted in weak staining (Fig. 4A). Three optical can also be observed in a phase contrast micrograph of the sections were taken sequentially through a fixed monolayer same field (Fig. 1D). Nonspecific binding of the secondary of astrocytes that had been double-labeled with anti-CK antibodies was not detectable with antimouse IgG-FITC and propidium iodide (PI).P I stains total nucleic acid and is
-
-
P. MANOS AND J. EDMOND
276
Fig. 1. Fluorescent localization of creatine kinase (CK), glial fibrillary acidic protein (GFAP)and nuclei in cultured astrocytes. Astrocytes were treated sequentially with the following antibodies (as indicated in Materials and Methods): goat anti-CK, donkey antigoat I&-LRh (A), mouse anti-GFAF', donkey antimouse I&-FITC (B).The final antibody solution contained the Hoechst nuclear stain (C). Corresponding field as seen with phase contrast optics (D).CK has a cytoplasmic staining
-
specific for DNA in RNase-treated samples. Each section represents a depth of focus of 0.7 pm (Wells et al., '891, and the images were recorded simultaneously in the same sections. CK staining is present in the cell nucleus in each optical section taken through the monolayer (Fig. 4B,1,2,3, left panel). CK positive staining in the nucleus appeared in
pattern that is distinct from GFAP and a nuclear staining pattern that co-localizes with the Hoechst stain. Photomicrographs were obtained from a conventional fluorescence microscope. To determine the extent of nonspecific binding of the secondary antibodies, control wells were incubated with 5%donkey serum instead of the primary antibodies and treated sequentially with antimouse I&-FITC (El and antigoat IgGLRh (F).Corresponding phase contrast field (GI. Bar : 25 pm.
the same sections through the nucleus as PI-labeled DNA (Fig. 4B 1,2,3,right panel), which demonstrates that CK is located within the nucleus. However, the labeling pattern of anti-CK and PI were clearly different in each nuclear section. Under conditions containing RNase, PI stains double-stranded DNA whereas CK staining has a particu-
CREATINE KINASE IN NUCLEAR ENERGY METABOLISM
Fig. 2. Fluorescent localization of creatine kinase (CK), tubulin and nuclei in cultured astrocytes. Astrocytes were sequentially incubated with the following antibodies (as indicated in Materials and Methods): rabbit anti-CK, donkey antirabbit I&-FITC (A), mouse antitubulin, donkey antimouse IgGLRh (B).The final antibody solution contained the Hoechst stain (C). Corresponding field as seen with phase contrast optics (D).These anti-CK antibodies, raised in rabbit, produce a similar
277
labeling pattern as that seen with the goat anti-CK antibodies (Fig. 1). Photomicrographs were obtained from a conventional fluorescence microscope. To determine the extent of nonspecific binding of the secondary antibodies, control wells were incubated with 5% donkey serum instead of the primary antibodies and treated with the secondary antibodies: antirabbit IgG-FITC (El and antimouse IgG-LRh (F). Corresponding phase contrast field ( G ) .Bar = 25 km.
278
P. MANOS AND J. EDMOND
Fig. 3. Removal of the pattern of CK staining by preadsorption of anti-CK serum with purified CK protein. Astrocytes were incubated with anti-CK serum (rabbit) that had been treated with either a CK-buffer (B) or with the purified CK protein (El as described in Materials and Methods. The cells were then incubated with antirabbit IgG-LRh containing the Hoechst stain. (A,D) Hoechst nuclear stain;
(C) is the correspondingphase contrast field as in (A) and (B); (F)is the corresponding phase contrast field as in (D) and (El.(A,B,C) were photographed, developed, and printed under identical conditions (film speed settings and exposure times) as (D,E,F). Photomicrographs were obtained from a standard fluorescence microscope. Bar = 50 km.
late, granular-likepattern throughout the nucleus, which is distinct from the pattern of staining for DNA. The pattern of CK staining was also compared with the staining pattern produced by labeling total nucleic acid (Fig. 5 ) . Dual channel images were recorded in a similar manner as described for Figure 4, except the cultures were not pretreated with RNase. PI therefore binds both doublestranded DNA and RNA. Six optical sections were taken through the cell monolayer. The first section was taken with the plane of focus just above the cell monolayer (Fig. 5A). The arrows indicate a representative astrocyte in which the CK stained nucleus (left panel) and PI-stained nucleic acids (right panel) can be very clearly seen (Fig. 5A,B) as the nucleus protrudes from the plane of the cell monolayer. P I staining is particularly intense in the subnuclear RNA-containing organelle, the nucleolus (Fig. 5C). Many of the nuclei contain more than one nucleolus. As successively deeper sections are taken through the monolayer, the amount of cytoplasmic cellular detail increases with both CK and PI staining (Fig. 5D,E). PI produces a distinct, filamentous pattern of staining. The nuclear CK and PI staining pattern in the cell indicated by the arrow becomes less apparent as sections are taken below the plane of focus that contains the nucleus (Fig. 5E,F). The pattern of CK staining in each section through the nucleus can be very clearly seen. CK is present throughout the nucleus and it does not stain the nucleolus. The particulate CK staining pattern is suggestive of a nucleoplasmic (karyoplasmic) distribution.
DISCUSSION Creatine kinase, an enzyme important in energy homeostasis, is distributed uniformly throughout the cytoplasm in cultured astrocytes and is especially intense in the nucleus. This finding of intense immunoreactive CK-B in the astrocyte nucleus was unexpected and, to our knowledge, has not been reported previously. The use of confocal microscopy has enabled us to elucidate the nuclear substructure of CK. CK has a punctate, granular-like pattern throughout the nucleus, which indicates that CK is a nucleoplasmic protein. CK is not found in the nucleolus and has a distinctly different staining pattern from the staining pattern produced by labeling either DNA or RNA. We believe the presence of intensely staining CK in the nucleus indicates that CK and the creatine-creatine phosphate energy pathway may provide, in part, the ATP necessary for the energy dependent processes of nuclear function. We used several immunological controls to assess the specificity of antibody binding. Nigg ('88) has suggested that irregular and nonspecific binding can be a potential problem, particularly with nuclear antigens, and can occur with a variety of immune or nonimmune sera. We found that two different antisera raised against the CK protein, in different species, produced a similar pattern of CK staining. The distribution of the CK label and the intense nuclear staining were nearly identical with both the goat and the rabbit antisera, even though different fluorochromes were conjugated to the second antibody (in Fig. lA, CK was
CREATINE KINASE IN NUCLEAR ENERGY METABOLISM
279
Fig. 4. Serial optical sections of immunostained CK and propidium iodide-stained DNA,using confocal microscopy. Control experiments in which astrocytes were treated with nonimmune rabbit serum instead of anti-CK and labeled with antirabbit IgG-FITC resulted in weak staining (A, left panel). Crossover of the FITC signal into the PI channel was not detectable. Bar = 25 bm (A,right panel). B.Astrocytes were treated with RNase before labeling with anti-CK serum (rabbit) and antirabbit
IgG-FITC (left panel). The second antibody solution contained PI (right panel). See Materials and Methods for more detail. B-1,2,3 represent three consecutive optical sections taken with a depth of focus of -0.7 pm and a distance of 0.5 ym along the Z axis. CK has a nucleoplasmic distribution that is distinct from DNA. Five scans were accumulated and averaged for each image. Bar = 10 pm.
detected with LRh; in Fig. 2A, CK was detected with FITC). In addition, the same pattern of CK staining was observed for the rabbit derived anti-CK with both FITC (Figs. 2A, 5) and LRh (Fig. 3B) as the fluorochrome. These controls eliminated the possibility of artifacts due to differences in binding of the secondary antibodies or intensity differences in the fluorochromes. Nonspecific binding of the fluoro-
chrome-conjugated secondary antibodies was also minimal (Figs. lE,F,G, 2E,F,G). We were also able to substantially decrease CK staining in the cytoplasm and nucleus by pre-treating the antiserum with the purified CK protein (Fig. 3). Nonspecific labeling with nonimmune rabbit serum was also minimal (Fig. 4A). These data argue convincingly that the labeling pattern produced with anti-CK is
280
P. MANOS AND J. EDMOND
-
Fig. 5. Serial optical sections of immunostained CK and propidium iodide stained total nucleic acid, using confocal microscopy. Astrocytes were labeled with anti-CK (rabbit) serum and antirabbit IgG-FITC (left panel) as described in Material and Methods. The second antibody solution contained PI (right panel). (A-F) Six consecutive optical
sections were taken with a depth of focus of 0.7 Fm and a distance of 1.0 km along the Z axis. The pattern of CK stainingis distinct from the pattern of staining for either RNA or DNA. Note the intense RNA stained nucleoli (right panels). Five scans were accumulated and averaged for each image. Bar = 25 km.
specific for CK and that CK is present in the nucleus of astrocytes cultured under these conditions. Several groups have examined the immunohistochemical localization of CK in brain sections (Yoshimine et al., '83; Kato et al., '86; Ikeda and Tomonaga, '87). CK immunoreactivity was found in the astrocyte cytoplasm but was not reported to stain the nucleus, nor was nuclear staining obvious from the photomicrographs presented. The lack of detectable CK in the astrocyte nucleus in the above studies may be attributable to the use of organic fixatives like methanol and ethanol. These fixatives have been reported
to extract or destroy nuclear antigens (Nigg, '88). In our experience, the use of methanol as a fixative greatly diminishes the intensity of total CK staining. Alternatively, the lack of detectable CK in the nucleus of astrocytes in brain sections may result from the fact that the sections were taken from adult brains, whereas the cultured astrocytes were obtained from neonatal brains. The localization of CK in the nucleus may result from the particular energy requirements of the nucleus during a specific period in development or in certain physiological states. In this regard, astrocytes in this study were reseeded 24 hours
CREATINE KINASE IN NUCLEAR ENERGY METABOLISM before preparation for immunocytochemistry. The replating of the cells may have an effect on the expression and/or subcellular localization of CK as the cells respond to the trauma of subculturing. Many enzymes within the nucleus require high energy phosphate in the form of (d)NTPs or, specifically,ATP. The (d)NTPs serve as substrates for both RNA and DNA directed polymerases in the synthesis of nucleic acids, and they are believed to provide energy for macromolecular transport (Newport and Forbes, '87). ATP is required, specifically, by Poly (A) polymerase and other nuclear enzymes such as ligases, topoisomerases, and helicases. (d)NTP's are generated by the addition of a phosphate to the corresponding NDP by cytoplasmic NDP kinases, using ATP as the phosphate donor. Cytoplasmic ATP can be generated from substrate level phosphorylation (anaerobic) or oxidative phosphorylation (aerobic). The nucleus, as an organelle, does not contain any primary energy generating mechanisms and therefore depends on the recruitment of ATP and other high energy phosphate compounds from the cytoplasm. Creatine phosphate, as an energy donor, may provide a local and sustained concentration of ATP at specific sites in the nucleus. Cellular ATPases are unable to use the high energy phosphate in creatine phosphate unless CK is present. Nuclear ATPases, therefore, need not compete with other cellular ATPases for ATP, as only ATPases associated with CK are able to use the high energy phosphate in creatine phosphate. Thus, CK and the creatinecreatine phosphate energy "shuttle" may form a link between the generation of energy by oxidative metabolism in the cytoplasm with substrate level phosphorylation in the astrocyte nucleus. Macromolecules are believed to be transported across the nuclear envelope via the nuclear pore by an active transport mechanism (Newport and Forbes, '86). The nuclear pore complex within the nuclear envelope is thought to be comprised of the machinery for the translocation of macromolecules, the contractile machinery, and the ATP hydrolyzing machinery. The export of messenger ribonucleoprotein particles (Agutter, '88) and ribosomes (Schumm et al., '79) and import of some nuclear proteins (Newmeyer, '90; Wagner et al., '90) has been shown to be energy dependent. ATPase activity has been detected histochemically on the nuclear envelope (Sikstrom et al., '76; Vorbrodt and Maul, '80; Fox et al., 'Bl), in discrete areas within the nucleolus (Fox et al., '81; Sikstrom et al., '761, with structures known to contain ribonucleoproteins (Vorbrodt and Maul, '801, and in deposits scattered throughout the nucleoplasm (Sikstrom et al., '76). In addition, ATPases are believed to be associated with nuclear contractile proteins (LeStourgeon, '78; Berrios and Fisher, '86). Thus nuclear CK may provide high energy phosphate in the region of these ATPases that are involved in macromolecular transport, contractile events, or other movements within the nucleus. The use of confocal microscopy as a detection method in these immunofluorescent studies has allowed us to elucidate structural details of CK in the astrocyte nucleus in greater detail than possible with conventional fluorescence microscopy. CK appears to have a ''soluble'' nucleoplasmic distribution. The pattern of CK staining in the nucleus was intriguing, and nonhomogeneous with the pattern of staining of either DNA or RNA. In this regard, a similar nucleoplasmic fluorescent staining pattern, which was incongruent with the staining pattern for DNA, has been observed for the product of the fos protooncogene (Nishiku-
281
ra and Murray, '87; Shuman et al., '89). We are very interested in the role CK serves in the astrocyte nucleus and in determining whether CK binds to DNA or interacts with other nuclear macromolecules.
ACKNOWLEDGMENTS We thank Justine Tseng for technical assistance during the preliminary stages of this work, Dr. Guy K. Bryan for critical review of the manuscript, and Drs. Michael S. Levine and Dorwin Birt for their assistance with the confocal microscope and computer imaging system. This work was supported by HD 06576.
LITERATURE CITED Agutter, P.S. (1988) Nucleo-cytoplasmic transport of mRNA Its relationship to RNA metabolism, subcellular structures and other nucleoplasmic exchanges. Prog. Mol. Subcell. Biol. 10:15-96. Austead, N., R.A. Korsak, J.W. Morrow, and J. Edmond (1991) Fatty acid oxidation and ketogenesis by astrorvtes in primary culture. J. Neurochem. 56:1376-1386. Berrios, M., and P.A. Fisher (1986) A myosin heavy chain-like polypeptide is associated with the nuclear envelope in higher eukaryotic cells. J. Cell Biol. 103:711-724. Bessman, S.P., and P.J. Geiger (1981) Transport of energy in muscle: The phosphorylcreatine shuttle. Science 211:448-452. Bessman, S.P., and C.L. Carpenter (1985) The creatine-creatine phosphate energy shuttle. Ann. Rev. Biochem. 54.832462. Bignami, A., L.F. Eng, A. Dahl, and C.T. Uyeda (1972) Localization of the glial fibrigary acidic protein in astrocytes by immunofluorescence. Brain Res. 43:429435. Debus, E., K. Weber, and M. Osborn (1983) Monoclonal antibodies specific for glial fibrillary acidic (GFA) protein and for each of the neurofilament triplet polypeptides. Differentiation 25: 193-203. Edmond, J., R.A. Robbins, J.D. Bergstrom, R.A. Cole, and J. deVellis (1987) Capacity for substrate utilization in oxidative metabolism by neurons, astrocytes and oligodendrocytes from developing brain in primary culture. J. Neurosci. Res. 18:551-561. Fox, N., C. Fernandez, and G.P. Studzinski (1981) Visualization of nucleolar substructure in cultured human fibroblasts by magnesium-activated adenosine triphosphate reaction. J. Histochem. Cytochem. 29: 11151120. Gray, J.W., and P. Coffin0 (1979) Cell cycle analysis by flow cytometry. 'Methods Enzymol58:233-248. Hawkins, R. (1985) Cerebral Energy Metabolism. In D.W. McCandless (ed): Cerebral Energy Metabolism and Metabolic Encephalopathy. New York: Plenum Press, pp. 3-23. Hertz, L. (1981) Features of astrocytic function apparently involved in the response of central nervous tissue to ischemia-hypoxia. J. Cereb. Blood Flow Metab. 1:143-153. Ikeda, K., and M. Tomonaga (1987) The presence of creatine kinase (CK)-immunoreactive neurons in the zona incerta and lateral hypothalamic areas of the mouse brain. Brain Res. 435348-350. Jones, K.H., and D.A. Kniss (1987) Propidium iodide as a nuclear counterstain for immunofluorescence studies on cells in culture. J. Histochem. Cytochem. 35:123-125. Kato, K., F. Suzuki, A. Shimizu, H. Shinohara, and R. Semba (1986) Highly sensitive immunoassay for rat brain-type creatine kinase: Determination in isolated Purkinje cells. J. Neurochem. 46:1783-1786. Kenyon, G.L., and G.H. Reed (1983) Creatine kinase: Structure-activity relationships. Adv. Enzymol. 54:367-426. LeStourgeon, W.M. (1978) The occurrence of contractile proteins in nuclei and their possible function. In H. Busch (ed):The Cell Nucleus. Vol. VI. New York: Academic Press, Inc, pp 305-326. Manos, P., G.K. Bryan, and J. Edmond (1991) Creatine kinase in postnatal rat brain development and in cultured neurons, astrocytes and oligodendrocytes. J. Neurochem. 56:2101-2107. Martinez-Hernandez, A,, K.P. Bell, and M.D. Norenberg (1977) Glutamine synthase: Glial localization in brain. Science 195:1356-1358. McCarthy, K.D., and J. de Vellis (1980) Preparation of separate astroglia and oligodendroglia cell cultures from rat cerebral tissue. J. Cell Biol. 85:890-902.
282 Meyer, R.A., H.L. Sweeney, and M.J. Kushmerick (1984) A simple analysis of the “phosphocreatine shuttle.” Am. J. Physiol. 15:C365-C389. Muller, W., and F. Gautier (1975) Interactions of heteroaromatic compounds with nucleic acids. A.T-specific non-intercalating DNA ligands. Eur. J. Biochem. 54:385-394. Newmeyer, D.D. (1990) Nuclear import in vitro. Prog. Mol. Subcell. Biol. IIr12-50. Newport, J.W., and D.J. Forbes (1987)The nucleus: Structure, function and dynamics. Ann. Rev. Biochem. 56t535-565. Nigg, E.A. (1988)Nuclear function and organization: The potential of immunochemical approaches. Int. Rev. Cytol. I IOr27-92. Nishikura, K., and J.M. Murray (1987) Antisense RNA of proto-oncogene c-fos blocks renewed growth of quiescent 3T3 cells. Mol. Cell Biol. 73339-649. Pawley, J.B. (1990)Handbook of Biological Confocal Microscopy. New York: Plenum Press, 232 pp. Schumm, D.E., M.A. Niemann, T. Palayoor, and T.E. Webb (1979) In vivo equivalence of a cell-free system from rat liver for ribosomal RNA processing and transport. J. Biol. Chem. 254r12126-12130. Shuman, H., J.M. Murray, and C. DiLullo (1989) Confocal microscopy: An overview. Biotechniques 7: 154-163. Sikstrom, R., J. Lanoix, and J.J.M. Bergeron (1976) An enzymatic analysis of a nuclear envelope fraction. Biochem. Biophys. Acta 448:88-102. Tombes, R.M., and B.M. Shapiro (1985) Metabolitechanneling: aphosphorylcreatine shuttle to mediate high energy phosphate transport between sperm mitochondrion and tail. Cell 42r325-334.
P. MANOS AND J. EDMOND Vorbrodt, A,, and G.G. Maul (1980) Cytochemical studies on the relation of nucleoside triphosphatase activity to ribonucleoproteins in isolated rat liver nuclei. J. Histochem. Cytochem. 28t27-35. Wagner, P., J. Kunz, A. Koller, and M.N. Hall (1990) Active transport into the nucleus. FEBS Letters 275:l-5. Wallimann, T., and H. Eppenberger (1990) The subcellular compartmentation of creatine kinase isoenzymes as a precondition for a proposed phosphoryl-creatine circuit. Prog. Clin. Biol. Res. 344:877-889. Wallimann, T., M.Wyss, D. Brdiczka, K. Nicolay, and H. Eppenberger (1992) Intracellular compartmentation, structure and function of creatine kinase isoenzymes in tissues with high and fluctuating energy demands: The “phosphocreatine circuit” for cellular energy homeostasis. Biochem. J. 281t21-40. Watts, D.C. (1973) Creatine kinase (adenosine 5’-triphosphate-~reatine phosphotransferase). The Enzymes 8:383-455. Wells, K.S., D.R. Sandison, J. Stricker, and W.W. Webb (1990) Quantitative fluorescence imaging with laser scanning confocal microscopy. In J. Pawley (ed): The Handbook of Biological Confocal Microscopy. New York: Plenum Press, pp. 27-39. White, J.G., W.B. Amos, and M. Fordham (1987) An evaluation of confocal versus conventional imaging of biological structures by fluorescence light microscopy. J. Cell Biol. 105t41-48. Wilson, T. (1990) Confocal Microscopy. New York: Academic Press, 426 pp. Yoshimine, T., K. Morimoto, H.A. Homburger, and T. Yanagihara (1983) Immunochemical localization of creatine kinase-BB isoenzyme in human brain: Comparison with tubulin and astroprotein. Brain Res. 265r101108.
