Journal of Neurocytology 21, 458-467 (1992)
Astroglial membrane structure is affected by agents that raise cyclic AMP and by phosphatidylcholine phospholipase C J. H . T A O - C H E N G
1., J. P. B R E S S L E R 2 a n d M . W . B R I G H T M A N ~
1 Laboratory of Neurobiology, National Institute of NeurologicaI Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892, USA 2 Kennedy Research Institute, Baltimore, MD 21205, USA Received 28 August 1991; revised 4 February 1992; accepted 10 February 1992
Summary The role of signal transduction mechanisms in the production of the characteristic orthogonal arrays of particle assemblies in the astroglial plasma membrane was investigated in vitro by freeze-fracture electron microscopy. Agents which raise cellular cAMP levels and subsequently activate protein kinase A, such as forskolin (50 b~M),isoproterenol (10 p~M)and 8-bromo-cAMP (1 raM), increased the density, the number of assemblies per unit area of cleaved cell membrane, and the frequency of astrocytes with assemblies. Agents that lead to the activation of protein kinase C, such as phorbo112,13-myristate acetate (at 50 riM) and choline-dependent phospholipase C (at 0.01-0.1 U ml 1), did not affect the assembly concentration. Thus, protein kinase A but not protein kinase C appears to be involved in the production of assemblies or their insertion into the astroglial plasma membrane. Although choline-dependent phospholipase C did not affect the astroglial assemblies, it caused the non-assembly, background particles to aggregate. A choline-dependent phospholipase C from a different source (B. cereus) was also active though at a higher concentration. Phospholipases of different specificities, such as phospholipase A2, phospholipase D or inositol-dependent phospholipase C were inactive over a wide range of concentrations. Two other astroglia derived cells, M~ller cells and cells of the C6 glioma ceil line, were also similarly affected by choline-dependent phospholipase C, while six other cells types including neurons, endothelial cells and fibroblasts were unaffected. It appears that phosphatidylcholine plays a significant role in determining the membrane structure of astrocytes. In a search for a means of isolating the assemblies, the binding of three lectins: ConA, WGA and PNA, conjugated to gold, was tested by label-fracture to ascertain whether the assemblies have an external oligosaccharide component. None of the lectins bound specifically to assemblies.
Introduction The function of the astroglial m e m b r a n e assemblies, orthogonal arrays of intramembranous particles, is u n k n o w n (Dermietzel, 1974; Landis & Reese, 1974). Brain endothelial cells and meningeal cells reproducibly induced a substantial increase in the density of the astroglial assemblies w h e n cells from the two types were co-cultured with astrocytes (Tao-Cheng et ai.,1990). The endothelial and meningeal cells m a y stimulate astrocytes to increase the assemblies either by secretory factors or by cell contact. Once the astrocytes are stimulated, second messages should be produced which regulate the synthesis of the assemblies. The purpose of the present experiments was to determine the signal transduction mechanisms that * To whom correspondence should be addressed. 0300M864/92 $03.00 +.12 9 1992 Chapman and Hall Ltd
control the production of assemblies. We chose to examine two well-defined mechanisms k n o w n to play a major role in the differentiation of several cell types (for review, see Gerisch, 1987). One system involves the activation of protein kinase A by cAMP. In this system an agonist activates protein kinase A by stimulating the production of cAMP t h r o u g h a receptor mediated mechanism, e.g., activation of beta adrenergic receptors with isoproterenol. The level of cAMP in cells can also be artificially increased by using a reagent that increases cAMP t h r o u g h a receptor i n d e p e n d e n t mechanism (e.g., forskolin) and a drug that inhibits the degradation of cAMP, for example, a phosphodiesterase inhibitor. Another transduction mechanism involves activation of protein kinase C by
Forskolin a n d P h o s p h o l i p a s e C affect astroglial m e m b r a n e s diacylglycerols. T h e s e lipids are g e n e r a t e d t h r o u g h a r e c e p t o r - m e d i a t e d m e c h a n i s m . Diacylglycerols are usually cleaved f r o m either p h o s p h a t i d y l c h o l i n e or p h o s p h a t i d y l i n o s i t o l b y a p p r o p r i a t e e n z y m e s such as choline- or i n o s i t o l - d e p e n d e n t p h o s p h o l i p a s e C. Protein kinase C can also be artificially stimulated b y u s i n g a p h o r b o l ester that w o r k s identically to diacylglycerols. In the quest of isolating the a s s e m b l y protein a n d m a k i n g an a n t i b o d y against it, w e h a v e also u s e d the technique of label-fracture (Pinto de Silva & Kan, 1984) to see if a n y lectin specifically b i n d s to a n y s u g a r m o i e t y that m i g h t be externally associated w i t h the assemblies. Preliminary results h a v e b e e n p u b l i s h e d in abstracts ( T a o - C h e n g et al., 1987a, 1988), a n d a r e v i e w (Brightman & T a o - C h e n g , 1990).
Materials and methods
Astroglial cultures Astroglial cultures were prepared as described previously (Tao-Cheng et aI., 1987b). Briefly, cerebral cortex tissue was collected from two-day-old rats and the meninges carefully removed. The tissue was mechanically dissociated by trituration with a i ml tuberculin syringe or by pressing and passing it through 130b~m and 74 ixm metal meshes. The dissociated tissues were spun down, resuspended and seeded with supplemented Basal Eagle's Medium plus 15% foetal calf serum for the first 3-4 days. After the initial seeding, the serum level was reduced to 10% and the culture media were changed twice a week. Secondary cultures (see McCarthy & de Vellis, 1980) were used in the present experiments for enrichment of astrocytes and for low baseline number of assemblies in these cultures (Tao-Cheng et al., 1990). Briefly, 7-10-day-old primary cultures in flasks were shaken at 200 rpm for 18-24 h on a rotary shaker to loosen the oligodendrocytes and other cell types. The cells were then thoroughly washed, trypsinized and seeded. For electron microscopy, cells were grown on Lux plastic cover slips and processed on the cover slips throughout the procedures. For immunocytochemical staining, some cells in each batch of the astrogIial culture were grown on glass cover slips and stained for glial fibrillary acidic protein (GFAP, Bignami et al., 1972) to assess the percentage of astrocytes in the preparation.
Agents that raise cAMP Astroglial cultures were incubated with (1) forskolin (Sigma) at 50 IxMfor 5 h to 5 days, (2) isoproterenol at 10 ~M for 5 h to 3 days, (3) isoproterenol with IBMX (3-isobutyl-l-methylxanthine, 1 txM), for 5 h to 3 days, (4) 8-bromo-cAMP at I mM for 5 h to 2 days, (5) control medium with 0.2% ethanol, the final concentration of ethanol that was included in the forskolin solution, for 5 h to 5 days, and (6) control medium without ethanol, for 5 h to 5 days. A rat glioma cell line, C6, the cells of which do not have assemblies, was cultured and treated with forskolin or
459
isoproterenol as previously described (Bressler & Tinsely, 1990) to see if assemblies could be induced.
Agents that activate protein kinase C(PKC) Astroglial cultures were treated with agents that affect PKC: (1) phorbol 12,13-myristate acetate (PMA), at 50riM for 5-7 h, and 1-5 days, and (2) choline-dependent phospholipase C(PC-PLC) from C. perfringens (Sigma) at 0.01-
0.1 Uml 1for2-7hand 1-3 days. The effect of choline dependent phospholipase C on non-assembly, background particles Astroglial cultures were incubated for 2 h, 5-7 h and 18-24 h with: (a) PC-PLC from C. perfringens at 0.01, 0.03 and 0.1 Um1-1, (b) the above PC-PLC without calcium in the solution, (c) PC-PLC from B. cereus (Sigma) at 0.1, 1 and 5 U ml -~. Astroglial cultures were also incubated with other phospholipases (PLCs) for 2-7 h, and 1-3 days: (a) phosphotidylinositol PLC (PI-PLC) from B. cereus (Boeringer Mannhelm) at 0.1 and 1 Um1-1. (b) phospholipase A2 (PL A2) (Sigma) at 1 U ml 1, (c) Phospholipase D (PLD) (Sigma) at 1 U ml 1. Guinea pig Mtiller cells (from R. Small, Small et al., 1991) and C6 cells were treated with PC-PLC from C. perfringens at 0.1 U m1-1 with or without calcium for 2-7 h to see if these close relatives of astrocytes are similarly affected by PC-PLC. Six control cell types were treated with PC-PLC at 0.1-1 U ml i to see if their membranes were affected similarly: (a) rat PC12 cells (from G. Gurrof), (b) rat cerebellar granule cells (from A. Novelli), (c) bovine brain endothelial cells (cf. Tao-Cheng et aI., 1987), (d) bovine aortic endothelial cells (from S. Williams), (e) rat fibroblasts (Tao-Cheng, 1987b), and (f) a fibroblast-like cell line from bovine trachea (ATCC, Rockville, MD).
Lectin binding Live cells were first labeled with colloidal gold conjugated ConA (concanavalin A, 30 fxgml 1), WGA (wheat germ aggtutinin, 10p~gml-1), and PNA (peanut agglutinin, 10 txg ml-1) (E.Y Lab) for 40 rain, washed in buffer, then fixed and processed for label-fracture (Pinto da Silva & Kan, 1984).
Freeze-fracture and thin section electron microscopy The cultures were fixed with 4% glutaraldehyde in 0.1M sodium cacodylate buffer at pH 7.4 for one to several hours and processed for freeze-fracture or thin section electron microscopy (Tao-Cheng et aI., 1990). For freeze-fracture, samples were washed in buffer, glycerinated through 10, 20 and 30% solutions in buffer, frozen in liquid freon 22 which was cooled by liquid nitrogen, fractured and replicated in a Balzers freeze-fracture unit. For label-fracture of the lectin labeled cells, the replicas were not cleaned with bleach, and only areas that exposed a single layer of E-face membrane were examined (Pinto da Silva & Kan, 1984). For thinsections, samples were washed, osmicated, mordanted with 1% uranyl acetate, dehydrated and embedded in EPON.
Morphometry for assembly density Assembly density is the number of assemblies per unit area of cleaved cell membrane. Random sampling of astrocytes was achieved by taking photomicrographs of assembly
460 containing membranes from different areas of the replicas (see methods of morphometry in Tao-Cheng et al., 1990). Because density of assembly varied among different batches of cultures, all the comparisons of assembly density were made on cells from the same batch. Furthermore, the assembly densities were so disparate in most cases that a simple average of the assembly values within the group may not have provided a truly representative profile of the distribution of the densities among the population of astrocytes. Therefore, the Smirnof test was chosen to compare the distributions of assembly densities rather than a test to compare the averages of the densities among different experimental conditions. In certain experiments, the presence or absence of assemblies was recorded for every membrane profile examined in order to assess the frequency of astrocytes with assemblies. A membrane profile was a piece of membrane in the replica with demarcated borders. It is possible that one cell will give rise to more than one membrane profile in the replica. Duplication of samples was reduced by taking samples from different quadrants of the grid so that these samples were sufficiently far apart from one another.
Results
Stimulating cells through protein kinase A The role of protein kinase A in the regulation of assemblies was examined in astrocytes treated with various agents that increase cAMP. A 50 b~M concentration of forskolin, which increase cAMP levels by activating adenyl cyclase, significantly augmented assembly density in astroglial cultures. In one experiment, no assemblies were detected in a control, nontreated culture (Fig. la) after extensive examination of, at least, more than 200 membrane profiles, while many cells in the sister culture (Fig. lb), treated with forskolin for 1 day, had assembly densities ranging from 1-17 p~m-2. The augmentation of assembly density by forskolin progressed with time. The same batch of astroglial cultures treated for 3 days with forskolin displayed assembly densities of 2-27 ~m 2 (Fig. ld), while nontreated cultures exhibited lower densities of no more than 2.5 ~m -2 (Fig. lc). This difference is illustrated in a pair of micrographs of sister cultures of astrocytes with (Fig. 2b) or without (Fig. 2a) forskolin. Forskolin also increased the density of the background particles (Fig. 2). The frequency of astrocytes with assemblies was also increased by forskolin. For example, in one matched pair, 15 of 96 membrane profiles (-16%) in the control group had assemblies, while 33 out of 74 membrane profiles (-45%) in the forskolin treated group displayed assemblies. It should be noted that at least 70-80% of the cells in these astroglial cultures were immunopositive for GFAP, i.e., 70-80% of the cells were astrocytes. This result indicates that not all astrocytes had assemblies in vitro, and that forskolin
TAO-CHENG, BRESSLER and BRIGHTMAN induced some of these assembly-free astrocytes to express assemblies in their membranes. Isoproterenol, an agent which raises cAMP levels through the beta adrenergic receptor, also enhanced assembly density, although to a much lesser degree than forskolin. In one experiment where an astroglial culture was treated for 18-24 h with 10 ~M isoproterenol, 9 out of 31 membrane profiles, or - 3 0 % , displayed assembly densities ranging from 0.1-4 i.~m-2, while nontreated cultures had no assemblies. Furthermore, the frequency of astrocytes with assemblies increased to - 4 3 % when I~M IBMX, a phosphodiesterase inhibitor which inhibits cAMP degradation, was added to the isoproterenol treatment. The response to isoproterenol was short-lived. An increase in assembly density was observed 5 h and 24 h after treatment, but it disappeared by 48 h. On the other hand, cells treated with forskolin continued to exhibit increased assembly densities by the fifth day of treatment. This observation was not surprising since the number of beta-adrenergic receptor decreases after prolonged treatment with a beta adrenergic ligand (Harden, 1983). Finally, 8-bromo-cAMP, which is another reagent that activates protein kinase A, also increased assembly density and the frequency of astrocytes with assemblies. It is emphasized that the augmentation of assembly density induced by these agents was most noticeable when the astroglial cultures had an initially low baseline number of assemblies. For example, when the baseline number in the control cultures was 0-4 assemblies per I.tm 2, all six trials with forskolin yielded a significant increase of assemblies. On the other hand, if the nontreated cultures initially displayed high assembly densities, for example, 5-15 assemblies per p.m 2, there were not a consistent further increase in the assembly density after treatment. Only one out of three trials of forskolin treatment in cultures with high assemblies yielded a significant increase. Assemblies are not expressed by transformed cell lines (Tao-Cheng et al., 1990). Such a cell line that could be induced to express assemblies could serve as source for the purification of the components that comprise the assembly. A C6 rat glioma cell line was chosen for that purpose. This cell line displays several characteristics that are similar to astrocytes and is also highly responsive to forskolin in raising its cellular cAMP. However, although there was a general increase of membrane particle density in the cell membrane of C6 glioma cells, there were no detectable assemblies in the cells treated for 5 h to 5 days with forskolin or isoproterenol.
Stimulating cells through protein kinase C Another type of signal transduction mechanism involves the protein kinase C family. In this mechanism,
(9 a. A ( A s t r o c y t e s )
o m m
E b. A + F O R S K O L I N
for 1 day
O ~-
10
Iii ,I,,,,ll
o ~
e-
,
-o
r
40
>
30
~
20
a e"
10
o
c.A
m
d. A + F O R S K O L I N ~-
for 3 days
%
0
o~
lo
t
It
~111
5
(ill
flI
10
I
J ~
15
20
,I
t
I
25
Oensi~y (#assemblies/.m0
Fig. 1. Histograms of astroglial assembly density (number per/a,m 2) in the same batch of cells as control (a,c) or forskolin treated (b,d) at 1 (a,b) and 3 days (c,d) of treatment. The distributions of assembly densities between control_ and forskolin treated samples were significantly different at both time points (a vs b, c vs d). Fig. 2. (a) Control astrocyte membrane had a low assembly (circles) density; (b) astrocyte treated with forskolin for 3 days contained many more assemblies. • 90 000.
462 diacylglycerol, generated by phospholipid hydrolysis, induces the translocation and activation of protein kinase C from the cytosol to the membrane. Protein kinase C, similarly to protein kinase A, is known to phosphorylate a diverse number of proteins. The role of protein kinase C in assembly expression was first examined by using phorbol 12,13-myristate acetate (PMA). This phorbol ester is capable of substituting for diacylglycerol in protein kinase C activation. Astroglial cultures were treated with PMA at 50rim for 5-7 h, and up to 5 days. Assembly densities, or the frequency of cells with assemblies did not change compared to control, nontreated cultures. The role of protein kinase C in assembly production was also examined by a different reagent. Here, the endogenous diacylglycerol levels were elevated by treating intact cells with choline-dependent phospholipase C. This enzyme cleaves phosphatidylcholine into phosphocholine and diacylglycerol. In an astroglial culture treated with 0.01 units of phospholipase C from C. perfringens for 2 h, assembly density appeared not to be significantly affected, but the nonassembly, background particles were aggregated as compared to the control (Figs 3 and 4). Aggregated background particles are clearly different from assemblies because the background particles are heterogeneous in size, generally larger and taller, and randomly packed, while the subunit particles of the assemblies are uniform in size, shorter in height, and orthogonally packed. Nevertheless, the aggregated background particles often flanked the assemblies and made the identification of assemblies very difficult. In a few examples, where the assemblies were still discernible, the assembly density was not significantly different from that of nontreated control. The aggregation of the background particles is further analyzed in the next section.
TAO-CHENG, BRESSLER and BRIGHTMAN
Effect of choline-dependent phospholipase C on background particles Background particles consistently became aggregated in astrocytes that had been incubated for 2 - 7 h in medium containing phospholipase C from C. perfringens (Figs 3 and 4). The lowest effective enzyme concentration to induce aggregation of background particles was 0.01 U m1-1 (Fig. 4) and this effect was dose dependent (cf. Fig. 5). No effect on the particles within the nuclear membrane was observed (Fig. 5). Phospholipase C-induced particle aggregation on astroglial plasma membrane was transient; the background particles became randomly redispersed after 24 h of treatment (Fig. 6) without changing the incubating medium. Particle aggregation was most likely due to the activity of phospholipase C and not a contaminant since particle aggregation was not observed when calcium was omitted from the reaction (Fig. 7). Calcium is required for the activity of this enzyme. Phospholipase C treatment did not significantly change the astroglial membrane's appearance in thin sections as shown in Fig. 8. In astrocytes as well as in other cell types, the outer leaflet of the cell membrane sometimes appeared lighter than the inner leaflet after phospholipase C treatment. This observation is in agreement with the report that phospholipase C treatment generally removed a dense material, which was prominent after lanthanum staining and which was associated with the outer leaflet of the cell membrane (Lesseps, 1967). The cytoskeletal elements in this astroglial process treated with phospholipase C were arranged as they were in other cell types in vitro (Bridgman et al., 1986) and as they were in nontreated controls, with actin filaments at the farthest periphery and the microtubules in the next domain. Abundant
Figs 3-7. Freeze-fractured astroglial membranes in vitro, x 100 000. Fig. 3. Control astrocytes. Secondary culture. Assemblies (circles) were randomly distributed. Fig. 4. Same batch of cells as above, treated with choline-dependent phospholipase C (PC-PLC) at 0.01 U ml -I for 2 h. The background particles redistributed into aggregates interspersed with particle-free areas. The assemblies (arrows) were flanked by many aggregated background particles. Fig. 5. Astrocyte at fourth passage, treated with choline-dependent phospholipase C at 0.1 U ml -~ for 7 h. The background particles of the plasma membrane became progressively more aggregated (arrows). The assemblies were no longer easily discernible. The nuclear membrane (N) was not affected by the enzyme treatment. Fig. 6. Treatment with choline-dependent phospholipase C at 0.1 U m1-1 without Ca2+ for 2 h. The astroglial membrane structure was not affected by the enzyme treatment. Assemblies are marked by circles. Fig. 7. Treatment with choline-dependent phospholipase C at 0.1 U m1-1 for 24 h without changing the medium. The background particles were no longer aggregated. Assemblies (circles) again were clearly discernible and randomly distributed. Fig. 8. Thin-section of an astrocytic process. Treatment with choline-dependent phospholipase C at 0.05 U ml -~ for 2 h. The cross-sectioned plasma membrane appeared normal in thin sections (inset enlarged at x 300 000). AC, actin filaments, MT, microtubules, IF, intermediate filaments, M, mitochondria. • 100 000.
464 intermediate filaments, characteristic of astrocytes, were the most centrally located, and were sometimes intermingled with the microtubules (Fig. 8). Choline-dependent phospholipase C from B. cereus also caused astroglial membrane particles to aggregate, but the effective dose (1 U ml 1) was 100 times higher than that of the choline-dependent phospholipase C from C. perfringens. The reason for this difference may be that the two enzymes have different accessibilities to the astroglial membrane because of their different structural configurations. Other phospholipases such as an inositol-dependent phospholipase C, phospholipase D and phospholipase A2 did not cause the astroglial membrane particles to aggregate. Thus, it appears that this astroglial membrane structural change is caused specifically by the action of a choline-dependent phospholipase C. A specialized astrocyte-like cell, the Mfiller cell of the retina, and a transformed astroglioma cell line, C6, were likewise affected by choline-dependent phospholipase C. The Mfiller cells from guinea pig retina, like the C6 glioma cells, did not have assemblies in vitro. However, their membranes showed distinctive particle aggregations and particle-free areas (Figs 9, 11) after 2 h of enzyme treatment. The enzyme was inactive if calcium was omitted from the treatment (Fig. 10). Membrane structure of six other cell types was also assessed at concentrations of up to 1-5 Um1-1 of choline-dependent phospholipase C. A neuron-like cell line, rat PC12 cells (Fig. 12), and rat cerebellar granule neurons were not affected. Other cell types that were not affected include fibroblasts from rat skin and from bovine trachea (Fig. 13), and endothelial cells from bovine brain (Fig. 14) and aorta.
Label-fracture of lectin binding As the gold particles were smaller than assemblies, binding of lectin-gold conjugate to a single assembly could have been recognized. However, the three
TAO-CHENG, BRESSLER and BRIGHTMAN representative lectins used in this study, ConA, WGA and PNA, were not bound to astroglial assemblies. The lectin-gold conjugates were, instead, randomly distributed in the astroglial membranes.
Discussion
Mechanisms which regulate assemblies Two different signal transduction mechanisms that may regulate assembly densities in astrocytes were investigated. These signals were cAMP that activates members of the protein kinase A family and diacylglycerol, which activates members of the protein kinase C family. We found that three drugs which increase cAMP by different mechanisms, forskolin, isoproterenol, and 8-bromo cAMP, increased assembly density and also increased the number of cells expressing assemblies. However, two different reagents that lead to activation of protein kinase C, PMA and phospholipase C, did not alter the assembly densities. Thus it appears that assembly density is regulated by protein kinase A and not by protein kinase C. There are two different mechanisms by which cAMP may act to change assembly density. In one mechanism, assembly precursors are converted to recognizable orthogonal assemblies by a phosphorylation reaction catalyzed by protein kinase A. These precursors or assembly subunits may either reside in the membrane, or in the cytosol. If the precursors are located in the cytosol, a phosphorylation step may be required to translocate the precursors to the membrane or to assemble them into recognizable orthogonal arrays. There are several examples of cytosolic proteins that became modified and attached to plasma membrane (Woodgett et al., 1987). In addition, membrane functions are influenced by phosphorylation. For example, cAMP dependent phosphorylation modifies both beta-adrenergic receptors (Harden,
Figs 9-10. Mfiller cell membranes. Fourth passage, 10 days in vitro, x 100 000. Fig. 9. Treatment with choline-dependent phospholipase C at 0.1 U m1-1 for 2 h. There were aggregates of intramembranous particles and particle-free areas. Fig. 10. The same treatment of choline-dependent phospholipase C except without Ca2+. The membrane particles were not affected. Fig. 11. C6, an astroglial tumor cell line, treated with choline-dependent phospholipase C at 0.1 U ml-~ for 2 h. Membrane particles were aggregated, x 100 000. Fig. 12. PC12, member of a pheochromocytoma cell lines, treated with choline-dependent phospholipase C at 0.1 U ml -I for 2 h. Membrane structure was not affected, x 100 000. Fig. 13. A fibroblast-like cell from a cell line derived from bovine trachea, treated with choline-dependent phospholipase C at 0.5 U m1-1 for 5 h. Membrane structure was not affected, x 100 000. Fig. 14. Brain endothelial (BE) cells, fifth passage, 10 days in culture, treated with choline-dependent phospholipase C at 0.5 U ml -~ for 2 h. Membrane structure was not affected. TJ, tight junction, x 100 000.
466 1983) and K + channels in astrocytes (Neary et aI., 1988). The concentration of assemblies at the site where extracellular K § is siphoned into the retinal glial cell, the Mtiller cell, has been suggested to signify that the assembly constitutes such K § channels (Newman, 1985). In the second mechanism, the phosphorylation induced by protein kinase A increases the production of new assemblies through protein synthesis. The synthesis of other proteins in astrocytes, for example, glutamine synthetase (Juurlink et al., 1981) and GFAP (Pollenz & McCarthy, 1986; Backhovens et al., 1987), are increased after cAMP levels are elevated. We favour this hypothesis of augmented synthesis for two reasons. First, new assemblies require protein synthesis (Anders & Brightman, 1982) and second, assembly density continues to increase in forskolinstimulated astrocytes 24 h after stimulation. Mechanisms dependent on the synthesis of new proteins usually requires hours, while mechanisms dependent on protein phosphorylation usually requires minutes. In the present in vitro experiments, the highest density of assemblies attainable following cAMP stimulation was - 3 0 pum, 2. It is emphasized that this number was only 7.5% that of the highest density (-400 txm 2, Anders & Brightman, 1979) in the most superficial layer of the glia limitans in vivo. Furthermore, astroglial cultures that already had high assembly densities cannot be stimulated by forskolin to produce significantly more assemblies. Thus, it appears that astrocytes, in vitro, have a limited capacity to respond to this signal transduction in the production or intercalation of assemblies. Although the density of assemblies appears to be modulated by cAMP, their distribution is not. When astrocytes were cocultured with brain endothelial cells, aggregates of assemblies with density approaching 400 txm-2 were found (Tao-Cheng et al., 1990). Since similar aggregates were not found in forskolintreated astrocytes, it is not likely that this aggregation was caused by activation of protein kinase A. One possible explanation is that there may be some factor(s) concentrated in the basal lamina, which covers the glia limitans in situ and is also present in patches in astroglial and brain endothelial cocultures. The factors in the basal lamina may hold the assemblies at high densities in the cell membrane directly apposed to basal lamina (Gotow, 1984). Assemblies can be clumped by protein denaturants, but such aggregation is distinct from that brought about by co-culturing astrocytes and brain endothelium. Guanidine and urea cause astroglial assemblies to coalesce into densely packed clumps which appear as large rafts of assemblies with many subunits (Anders & Brightman, 1982), whereas the co-culturing brings about the aggregation of many regular-sized assemblies (Tao-Cheng et al., 1990).
TAO-CHENG, BRESSLER and BRIGHTMAN Non-assembly, background particle aggregation The astroglial, non-assembly, background particles aggregated after choline-dependent phospholipase C treatment. Other phospholipases with different specificities, such as inositol-dependent phospholipase C and phospholipase A2 or D, were inactive even at concentrations that were 10-100 times higher than the choline-dependent phospholipase C. These results suggest the following two conclusions. First, if it is assumed that each enzyme is equally reactive with the astroglial membrane, then the aggregation of background particles in astrocytes is caused by the enzymatic cleaving of a distinct phospholipid, since each enzyme exhibits a restricted specificity. Phosphatidylcholine appears to be the pivotal lipid for integrity of the astroglial plasma membrane since background particles only aggregated after treatment with cholinedependent PLC and not with inositol-dependent PLC. Second, the loss of arachidonic acid (the product of phospholipase A2), from a phospholipid did not induce aggregation of background particles. The loss of arachidonic acid leads to the production of a lysophospholipid, e.g., lysophosphatidylcholine, which, once produced, is not depleted from the membrane because it is quickly reacetylated at the membrane. In contrast, phosphatidylcholine hydrolysis by choline-dependent phospholipase C leads to the depletion of phosphatidylcholine from the membrane, and a longer time is needed for the production of new phosphatidylcholine as replacement. The depletion of this lipid may disrupt the balance of polar and apolar forces at the membrane and allow the background particles to aggregate. At 24 h, the astroglial membrane structure recovered to its normal morphology probably because more phosphatidylcholine was made in that time. Cells of astroglial origin were much more sensitive to phospholipase C treatment than other cell types, including neurons, endothelial cells and fibroblasts. This specific sensitivity may be related to the quantity of phosphatidylcholine or to a different role this phospholipid plays in membrane function in the astrocytes. Phosphatidylcholine may be divided into two pools. One pool is metabolized quickly and has a metabolic function, while the other pool is metabolized more slowly and serves a structural role. It appears that in astrocytes phosphatidylcholine exhibits a more important role in sustaining structure than in other cell types. Conclusion Astroglial assembly production can be augmented by cAMP but not by phorbol esters or diacylglycerols. Thus, protein kinase A but not protein kinase C is involved in assembly production. The assemblies appear to be the product of protein synthesis and not
Forskolin a n d P h o s p h o l i p a s e C affect astroglial m e m b r a n e s p h o s p h o r y l a t i o n since an increase in a s s e m b l i e s w a s still o b s e r v e d 4 8 h after stimulation. In a d d i t i o n , b a c k g r o u n d particles in a s t r o c y t e s a p p e a r u n i q u e l y sensitive to c h a n g e s in p h o s p h a t i d y l c h o l i n e levels. Therefore, this p h o s p h o l i p i d m a y exhibit a different role in a s t r o c y t e s c o m p a r e d to o t h e r cell types.
467
Acknowledgement W e t h a n k Dr K. P e t t i g r e w for advice o n statistical analysis, Dr R. Small for g u i n e a pig M/iller cells, Dr G. G u r o f f for PC12 cells, Dr A. Novelli for rat cerebellar g r a n u l e cell cultures, a n d Dr S. Williams for b o v i n e aortic e n d o t h e l i a l cells.
References ANDERS, J. I. & BRIGHTMAN, M. W. (1979) Assemblies of particles in the cell membranes of developing, mature and reactive astrocytes. Journal of Neurocytology 8, 777-95 ANDERS, J. J. & BRIGHTMAN, M. W. (1982) Particle assemblies in astrocytic plasma membrane are rearranged by various agents in vitro and cold injury in vivo. Journal of Neurocytology 11, 1009-29. BACKHOVENS, H., GHEUENS, J. & SLEGERS, H. (1987) Expression of gial fibrillary acidic protein in rat C6 glioma relates to vimentin and is independent of cell-cell contact. Journal of Neurochemistry 49, 348-54. BIGNAMI, A., ENG, L. F., DAHL, D. & UYEDA, T. (1972) Localization of glial fibrillary acidic protein in astrocytes by immunofluorescence. Brain Research 43, 429-35. BRESSLER, J. P. & TINSELY, P. (1990) Regulation of cAMP levels by protein kinase C in C6 rat glioma cells. Journal of Neuroscience Research 25, 81-6. BRIDGMAN, P. C., KACHAR, B. & REESE, T. S. (1986) The structure of cytoplasm in directly frozen cultured cells. II. Cytoplasmic domains associated with organelle movements., Journal of Cell Biology 102, 1510-21. BRIGHTMAN, M. W. & TAO-CHENG, J.-H. (1990) Mutually imposed structural changes in plasma membranes of astroglia and brain endothelium. In Differentiation and Functions of GliaI Cells (edited by LEVI, G.) pp. 107-15. New York: Alan R. Liss. DERMIETZEL, R. (1974) Junctions in the central nervous system of the cat. III. Gap junctions and membraneassociated orthogonal particle complexes (MOPC) in astrocytic membranes. Cell and Tissue Research 149, 121-35. GERISCH, G. (1987) Cyclic AMP and other signals controlling cell development and differentiation in Dictyostelium. Annual Review of Biochemistry 56, 853-79. GOTOW, T. (1984) Cytochemical characteristics of astrocytic plasma membranes specialized with numerous orthogonal arrays. Journal of Neurocytology 13, 43148. HARDEN, T. K. (1983) Agonist-induced desensitization of the beta-adrenergic receptor-linked adenylate cyclase. Pharmacological Reviews 35, 5-32. JUURLINK, B. H., SCHOUSBOE, A., JORGENSEN, O, & HERTZ, L. (1981) Induction by hydrocortisone of glutamine synthetase in mouse primary astrocyte cultures. Journal of Neurochemistry 36, 13642.
LANDIS, D. M. D. & REESE, T. S. (1974) Arrays of particles in freeze-fractured astrocytic membranes. Journal of Cell Biology 60, 316-20. LESSEPS, R. J. (1967) The removal of phospholipase C of a layer of lanthanum-staining material external to the cell membrane in embryonic chick cells. Journal of Cell Biology 34, 173-83. MCCARTHY, M. D. & DE VELLIS, J. (1980) Preparation of separate astroglial and oligodendroglial cell cultures from rat cerebral tissue. Journal of Cell Biology 85, 890-902. NEARY, J. T., NORENBERG, L. O. B. & NORENBERG, M. D. (1988) Protein kinase C in primary astrocyte cultures: cytoplasmic localization and translocation by a phorbol ester. Journal of Neurochemistry 50, 1179-84. NEWMAN, A. E. (1985) Membrane physiology of retinal glial (Mtiller) cells. Journal of Neuroscience 5, 2225-39. PINTO DA SILVA, P. & KAN, F. W. K. (1984) Label-fracture: a method for high resolution labeling of cell surfaces. Journal of Cell Biology 99, 1156-61. POLLENZ, R. S. & MCCARTHY, K. D. (1986)Analysis of cyclic AMP-dependent changes in intermediate filament protein phosphorylation and cell morphology in cultured astroglia. Journal of Neurochemistry 47, 9-17. SMALL, R. K., PATEL, P. & WATKINS, B. A. (1991) Response of Miiller cells to growth factors alters with time in cultme. Glia 4, 469-83. TAO-CHENG, J,-H., BRESSLER, I. P. & BRIGHTMAN, M. W. (1987a) Structural changes of astroglia membranes in vitro. Society for Neuroscience Abstracts 13, 120. TAO-CHENG, J.-H., NAGY, Z. & BRIGHTMAN, M. W. (1987b) Tight junctions of brain endothelium in vitro are enhanced by astrocytes. Journal of Neuroscience 7, 3293-9. TAO-CHENG, J.-H., BRESSLER, J. P. & BRIGHTMAN, M. W. (1988) Characterization of astroglial membrane assemblies by agents which activate protein kinase C. Society for Neuroscience Abstracts 14, 430. TAO-CHENG, J.-H., NAGY, Z. & BRIGHTMAN, M. W. (1990) Astrocytic orthogonal arrays of intramembranous particle assemblies are modulated by brain endothelial cells in vitro. Journal of Neurocytology 19, 143-55. WOODGETT, J. R., HUNTER, T. & GOULD, K. L. (1987) Protein kinase C and its role in cell growth. In Cell Membranes Methods and Reviews Vol. 3 (edited by ELSON, E., FRAZIER, W. & GLASER, L.) pp. 215--340. New York: Plenum Press.
Astroglial membrane structure is affected by agents that raise cyclic AMP and by phosphatidylcholine phospholipase C J. H . T A O - C H E N G
1., J. P. B R E S S L E R 2 a n d M . W . B R I G H T M A N ~
1 Laboratory of Neurobiology, National Institute of NeurologicaI Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892, USA 2 Kennedy Research Institute, Baltimore, MD 21205, USA Received 28 August 1991; revised 4 February 1992; accepted 10 February 1992
Summary The role of signal transduction mechanisms in the production of the characteristic orthogonal arrays of particle assemblies in the astroglial plasma membrane was investigated in vitro by freeze-fracture electron microscopy. Agents which raise cellular cAMP levels and subsequently activate protein kinase A, such as forskolin (50 b~M),isoproterenol (10 p~M)and 8-bromo-cAMP (1 raM), increased the density, the number of assemblies per unit area of cleaved cell membrane, and the frequency of astrocytes with assemblies. Agents that lead to the activation of protein kinase C, such as phorbo112,13-myristate acetate (at 50 riM) and choline-dependent phospholipase C (at 0.01-0.1 U ml 1), did not affect the assembly concentration. Thus, protein kinase A but not protein kinase C appears to be involved in the production of assemblies or their insertion into the astroglial plasma membrane. Although choline-dependent phospholipase C did not affect the astroglial assemblies, it caused the non-assembly, background particles to aggregate. A choline-dependent phospholipase C from a different source (B. cereus) was also active though at a higher concentration. Phospholipases of different specificities, such as phospholipase A2, phospholipase D or inositol-dependent phospholipase C were inactive over a wide range of concentrations. Two other astroglia derived cells, M~ller cells and cells of the C6 glioma ceil line, were also similarly affected by choline-dependent phospholipase C, while six other cells types including neurons, endothelial cells and fibroblasts were unaffected. It appears that phosphatidylcholine plays a significant role in determining the membrane structure of astrocytes. In a search for a means of isolating the assemblies, the binding of three lectins: ConA, WGA and PNA, conjugated to gold, was tested by label-fracture to ascertain whether the assemblies have an external oligosaccharide component. None of the lectins bound specifically to assemblies.
Introduction The function of the astroglial m e m b r a n e assemblies, orthogonal arrays of intramembranous particles, is u n k n o w n (Dermietzel, 1974; Landis & Reese, 1974). Brain endothelial cells and meningeal cells reproducibly induced a substantial increase in the density of the astroglial assemblies w h e n cells from the two types were co-cultured with astrocytes (Tao-Cheng et ai.,1990). The endothelial and meningeal cells m a y stimulate astrocytes to increase the assemblies either by secretory factors or by cell contact. Once the astrocytes are stimulated, second messages should be produced which regulate the synthesis of the assemblies. The purpose of the present experiments was to determine the signal transduction mechanisms that * To whom correspondence should be addressed. 0300M864/92 $03.00 +.12 9 1992 Chapman and Hall Ltd
control the production of assemblies. We chose to examine two well-defined mechanisms k n o w n to play a major role in the differentiation of several cell types (for review, see Gerisch, 1987). One system involves the activation of protein kinase A by cAMP. In this system an agonist activates protein kinase A by stimulating the production of cAMP t h r o u g h a receptor mediated mechanism, e.g., activation of beta adrenergic receptors with isoproterenol. The level of cAMP in cells can also be artificially increased by using a reagent that increases cAMP t h r o u g h a receptor i n d e p e n d e n t mechanism (e.g., forskolin) and a drug that inhibits the degradation of cAMP, for example, a phosphodiesterase inhibitor. Another transduction mechanism involves activation of protein kinase C by
Forskolin a n d P h o s p h o l i p a s e C affect astroglial m e m b r a n e s diacylglycerols. T h e s e lipids are g e n e r a t e d t h r o u g h a r e c e p t o r - m e d i a t e d m e c h a n i s m . Diacylglycerols are usually cleaved f r o m either p h o s p h a t i d y l c h o l i n e or p h o s p h a t i d y l i n o s i t o l b y a p p r o p r i a t e e n z y m e s such as choline- or i n o s i t o l - d e p e n d e n t p h o s p h o l i p a s e C. Protein kinase C can also be artificially stimulated b y u s i n g a p h o r b o l ester that w o r k s identically to diacylglycerols. In the quest of isolating the a s s e m b l y protein a n d m a k i n g an a n t i b o d y against it, w e h a v e also u s e d the technique of label-fracture (Pinto de Silva & Kan, 1984) to see if a n y lectin specifically b i n d s to a n y s u g a r m o i e t y that m i g h t be externally associated w i t h the assemblies. Preliminary results h a v e b e e n p u b l i s h e d in abstracts ( T a o - C h e n g et al., 1987a, 1988), a n d a r e v i e w (Brightman & T a o - C h e n g , 1990).
Materials and methods
Astroglial cultures Astroglial cultures were prepared as described previously (Tao-Cheng et aI., 1987b). Briefly, cerebral cortex tissue was collected from two-day-old rats and the meninges carefully removed. The tissue was mechanically dissociated by trituration with a i ml tuberculin syringe or by pressing and passing it through 130b~m and 74 ixm metal meshes. The dissociated tissues were spun down, resuspended and seeded with supplemented Basal Eagle's Medium plus 15% foetal calf serum for the first 3-4 days. After the initial seeding, the serum level was reduced to 10% and the culture media were changed twice a week. Secondary cultures (see McCarthy & de Vellis, 1980) were used in the present experiments for enrichment of astrocytes and for low baseline number of assemblies in these cultures (Tao-Cheng et al., 1990). Briefly, 7-10-day-old primary cultures in flasks were shaken at 200 rpm for 18-24 h on a rotary shaker to loosen the oligodendrocytes and other cell types. The cells were then thoroughly washed, trypsinized and seeded. For electron microscopy, cells were grown on Lux plastic cover slips and processed on the cover slips throughout the procedures. For immunocytochemical staining, some cells in each batch of the astrogIial culture were grown on glass cover slips and stained for glial fibrillary acidic protein (GFAP, Bignami et al., 1972) to assess the percentage of astrocytes in the preparation.
Agents that raise cAMP Astroglial cultures were incubated with (1) forskolin (Sigma) at 50 IxMfor 5 h to 5 days, (2) isoproterenol at 10 ~M for 5 h to 3 days, (3) isoproterenol with IBMX (3-isobutyl-l-methylxanthine, 1 txM), for 5 h to 3 days, (4) 8-bromo-cAMP at I mM for 5 h to 2 days, (5) control medium with 0.2% ethanol, the final concentration of ethanol that was included in the forskolin solution, for 5 h to 5 days, and (6) control medium without ethanol, for 5 h to 5 days. A rat glioma cell line, C6, the cells of which do not have assemblies, was cultured and treated with forskolin or
459
isoproterenol as previously described (Bressler & Tinsely, 1990) to see if assemblies could be induced.
Agents that activate protein kinase C(PKC) Astroglial cultures were treated with agents that affect PKC: (1) phorbol 12,13-myristate acetate (PMA), at 50riM for 5-7 h, and 1-5 days, and (2) choline-dependent phospholipase C(PC-PLC) from C. perfringens (Sigma) at 0.01-
0.1 Uml 1for2-7hand 1-3 days. The effect of choline dependent phospholipase C on non-assembly, background particles Astroglial cultures were incubated for 2 h, 5-7 h and 18-24 h with: (a) PC-PLC from C. perfringens at 0.01, 0.03 and 0.1 Um1-1, (b) the above PC-PLC without calcium in the solution, (c) PC-PLC from B. cereus (Sigma) at 0.1, 1 and 5 U ml -~. Astroglial cultures were also incubated with other phospholipases (PLCs) for 2-7 h, and 1-3 days: (a) phosphotidylinositol PLC (PI-PLC) from B. cereus (Boeringer Mannhelm) at 0.1 and 1 Um1-1. (b) phospholipase A2 (PL A2) (Sigma) at 1 U ml 1, (c) Phospholipase D (PLD) (Sigma) at 1 U ml 1. Guinea pig Mtiller cells (from R. Small, Small et al., 1991) and C6 cells were treated with PC-PLC from C. perfringens at 0.1 U m1-1 with or without calcium for 2-7 h to see if these close relatives of astrocytes are similarly affected by PC-PLC. Six control cell types were treated with PC-PLC at 0.1-1 U ml i to see if their membranes were affected similarly: (a) rat PC12 cells (from G. Gurrof), (b) rat cerebellar granule cells (from A. Novelli), (c) bovine brain endothelial cells (cf. Tao-Cheng et aI., 1987), (d) bovine aortic endothelial cells (from S. Williams), (e) rat fibroblasts (Tao-Cheng, 1987b), and (f) a fibroblast-like cell line from bovine trachea (ATCC, Rockville, MD).
Lectin binding Live cells were first labeled with colloidal gold conjugated ConA (concanavalin A, 30 fxgml 1), WGA (wheat germ aggtutinin, 10p~gml-1), and PNA (peanut agglutinin, 10 txg ml-1) (E.Y Lab) for 40 rain, washed in buffer, then fixed and processed for label-fracture (Pinto da Silva & Kan, 1984).
Freeze-fracture and thin section electron microscopy The cultures were fixed with 4% glutaraldehyde in 0.1M sodium cacodylate buffer at pH 7.4 for one to several hours and processed for freeze-fracture or thin section electron microscopy (Tao-Cheng et aI., 1990). For freeze-fracture, samples were washed in buffer, glycerinated through 10, 20 and 30% solutions in buffer, frozen in liquid freon 22 which was cooled by liquid nitrogen, fractured and replicated in a Balzers freeze-fracture unit. For label-fracture of the lectin labeled cells, the replicas were not cleaned with bleach, and only areas that exposed a single layer of E-face membrane were examined (Pinto da Silva & Kan, 1984). For thinsections, samples were washed, osmicated, mordanted with 1% uranyl acetate, dehydrated and embedded in EPON.
Morphometry for assembly density Assembly density is the number of assemblies per unit area of cleaved cell membrane. Random sampling of astrocytes was achieved by taking photomicrographs of assembly
460 containing membranes from different areas of the replicas (see methods of morphometry in Tao-Cheng et al., 1990). Because density of assembly varied among different batches of cultures, all the comparisons of assembly density were made on cells from the same batch. Furthermore, the assembly densities were so disparate in most cases that a simple average of the assembly values within the group may not have provided a truly representative profile of the distribution of the densities among the population of astrocytes. Therefore, the Smirnof test was chosen to compare the distributions of assembly densities rather than a test to compare the averages of the densities among different experimental conditions. In certain experiments, the presence or absence of assemblies was recorded for every membrane profile examined in order to assess the frequency of astrocytes with assemblies. A membrane profile was a piece of membrane in the replica with demarcated borders. It is possible that one cell will give rise to more than one membrane profile in the replica. Duplication of samples was reduced by taking samples from different quadrants of the grid so that these samples were sufficiently far apart from one another.
Results
Stimulating cells through protein kinase A The role of protein kinase A in the regulation of assemblies was examined in astrocytes treated with various agents that increase cAMP. A 50 b~M concentration of forskolin, which increase cAMP levels by activating adenyl cyclase, significantly augmented assembly density in astroglial cultures. In one experiment, no assemblies were detected in a control, nontreated culture (Fig. la) after extensive examination of, at least, more than 200 membrane profiles, while many cells in the sister culture (Fig. lb), treated with forskolin for 1 day, had assembly densities ranging from 1-17 p~m-2. The augmentation of assembly density by forskolin progressed with time. The same batch of astroglial cultures treated for 3 days with forskolin displayed assembly densities of 2-27 ~m 2 (Fig. ld), while nontreated cultures exhibited lower densities of no more than 2.5 ~m -2 (Fig. lc). This difference is illustrated in a pair of micrographs of sister cultures of astrocytes with (Fig. 2b) or without (Fig. 2a) forskolin. Forskolin also increased the density of the background particles (Fig. 2). The frequency of astrocytes with assemblies was also increased by forskolin. For example, in one matched pair, 15 of 96 membrane profiles (-16%) in the control group had assemblies, while 33 out of 74 membrane profiles (-45%) in the forskolin treated group displayed assemblies. It should be noted that at least 70-80% of the cells in these astroglial cultures were immunopositive for GFAP, i.e., 70-80% of the cells were astrocytes. This result indicates that not all astrocytes had assemblies in vitro, and that forskolin
TAO-CHENG, BRESSLER and BRIGHTMAN induced some of these assembly-free astrocytes to express assemblies in their membranes. Isoproterenol, an agent which raises cAMP levels through the beta adrenergic receptor, also enhanced assembly density, although to a much lesser degree than forskolin. In one experiment where an astroglial culture was treated for 18-24 h with 10 ~M isoproterenol, 9 out of 31 membrane profiles, or - 3 0 % , displayed assembly densities ranging from 0.1-4 i.~m-2, while nontreated cultures had no assemblies. Furthermore, the frequency of astrocytes with assemblies increased to - 4 3 % when I~M IBMX, a phosphodiesterase inhibitor which inhibits cAMP degradation, was added to the isoproterenol treatment. The response to isoproterenol was short-lived. An increase in assembly density was observed 5 h and 24 h after treatment, but it disappeared by 48 h. On the other hand, cells treated with forskolin continued to exhibit increased assembly densities by the fifth day of treatment. This observation was not surprising since the number of beta-adrenergic receptor decreases after prolonged treatment with a beta adrenergic ligand (Harden, 1983). Finally, 8-bromo-cAMP, which is another reagent that activates protein kinase A, also increased assembly density and the frequency of astrocytes with assemblies. It is emphasized that the augmentation of assembly density induced by these agents was most noticeable when the astroglial cultures had an initially low baseline number of assemblies. For example, when the baseline number in the control cultures was 0-4 assemblies per I.tm 2, all six trials with forskolin yielded a significant increase of assemblies. On the other hand, if the nontreated cultures initially displayed high assembly densities, for example, 5-15 assemblies per p.m 2, there were not a consistent further increase in the assembly density after treatment. Only one out of three trials of forskolin treatment in cultures with high assemblies yielded a significant increase. Assemblies are not expressed by transformed cell lines (Tao-Cheng et al., 1990). Such a cell line that could be induced to express assemblies could serve as source for the purification of the components that comprise the assembly. A C6 rat glioma cell line was chosen for that purpose. This cell line displays several characteristics that are similar to astrocytes and is also highly responsive to forskolin in raising its cellular cAMP. However, although there was a general increase of membrane particle density in the cell membrane of C6 glioma cells, there were no detectable assemblies in the cells treated for 5 h to 5 days with forskolin or isoproterenol.
Stimulating cells through protein kinase C Another type of signal transduction mechanism involves the protein kinase C family. In this mechanism,
(9 a. A ( A s t r o c y t e s )
o m m
E b. A + F O R S K O L I N
for 1 day
O ~-
10
Iii ,I,,,,ll
o ~
e-
,
-o
r
40
>
30
~
20
a e"
10
o
c.A
m
d. A + F O R S K O L I N ~-
for 3 days
%
0
o~
lo
t
It
~111
5
(ill
flI
10
I
J ~
15
20
,I
t
I
25
Oensi~y (#assemblies/.m0
Fig. 1. Histograms of astroglial assembly density (number per/a,m 2) in the same batch of cells as control (a,c) or forskolin treated (b,d) at 1 (a,b) and 3 days (c,d) of treatment. The distributions of assembly densities between control_ and forskolin treated samples were significantly different at both time points (a vs b, c vs d). Fig. 2. (a) Control astrocyte membrane had a low assembly (circles) density; (b) astrocyte treated with forskolin for 3 days contained many more assemblies. • 90 000.
462 diacylglycerol, generated by phospholipid hydrolysis, induces the translocation and activation of protein kinase C from the cytosol to the membrane. Protein kinase C, similarly to protein kinase A, is known to phosphorylate a diverse number of proteins. The role of protein kinase C in assembly expression was first examined by using phorbol 12,13-myristate acetate (PMA). This phorbol ester is capable of substituting for diacylglycerol in protein kinase C activation. Astroglial cultures were treated with PMA at 50rim for 5-7 h, and up to 5 days. Assembly densities, or the frequency of cells with assemblies did not change compared to control, nontreated cultures. The role of protein kinase C in assembly production was also examined by a different reagent. Here, the endogenous diacylglycerol levels were elevated by treating intact cells with choline-dependent phospholipase C. This enzyme cleaves phosphatidylcholine into phosphocholine and diacylglycerol. In an astroglial culture treated with 0.01 units of phospholipase C from C. perfringens for 2 h, assembly density appeared not to be significantly affected, but the nonassembly, background particles were aggregated as compared to the control (Figs 3 and 4). Aggregated background particles are clearly different from assemblies because the background particles are heterogeneous in size, generally larger and taller, and randomly packed, while the subunit particles of the assemblies are uniform in size, shorter in height, and orthogonally packed. Nevertheless, the aggregated background particles often flanked the assemblies and made the identification of assemblies very difficult. In a few examples, where the assemblies were still discernible, the assembly density was not significantly different from that of nontreated control. The aggregation of the background particles is further analyzed in the next section.
TAO-CHENG, BRESSLER and BRIGHTMAN
Effect of choline-dependent phospholipase C on background particles Background particles consistently became aggregated in astrocytes that had been incubated for 2 - 7 h in medium containing phospholipase C from C. perfringens (Figs 3 and 4). The lowest effective enzyme concentration to induce aggregation of background particles was 0.01 U m1-1 (Fig. 4) and this effect was dose dependent (cf. Fig. 5). No effect on the particles within the nuclear membrane was observed (Fig. 5). Phospholipase C-induced particle aggregation on astroglial plasma membrane was transient; the background particles became randomly redispersed after 24 h of treatment (Fig. 6) without changing the incubating medium. Particle aggregation was most likely due to the activity of phospholipase C and not a contaminant since particle aggregation was not observed when calcium was omitted from the reaction (Fig. 7). Calcium is required for the activity of this enzyme. Phospholipase C treatment did not significantly change the astroglial membrane's appearance in thin sections as shown in Fig. 8. In astrocytes as well as in other cell types, the outer leaflet of the cell membrane sometimes appeared lighter than the inner leaflet after phospholipase C treatment. This observation is in agreement with the report that phospholipase C treatment generally removed a dense material, which was prominent after lanthanum staining and which was associated with the outer leaflet of the cell membrane (Lesseps, 1967). The cytoskeletal elements in this astroglial process treated with phospholipase C were arranged as they were in other cell types in vitro (Bridgman et al., 1986) and as they were in nontreated controls, with actin filaments at the farthest periphery and the microtubules in the next domain. Abundant
Figs 3-7. Freeze-fractured astroglial membranes in vitro, x 100 000. Fig. 3. Control astrocytes. Secondary culture. Assemblies (circles) were randomly distributed. Fig. 4. Same batch of cells as above, treated with choline-dependent phospholipase C (PC-PLC) at 0.01 U ml -I for 2 h. The background particles redistributed into aggregates interspersed with particle-free areas. The assemblies (arrows) were flanked by many aggregated background particles. Fig. 5. Astrocyte at fourth passage, treated with choline-dependent phospholipase C at 0.1 U ml -~ for 7 h. The background particles of the plasma membrane became progressively more aggregated (arrows). The assemblies were no longer easily discernible. The nuclear membrane (N) was not affected by the enzyme treatment. Fig. 6. Treatment with choline-dependent phospholipase C at 0.1 U m1-1 without Ca2+ for 2 h. The astroglial membrane structure was not affected by the enzyme treatment. Assemblies are marked by circles. Fig. 7. Treatment with choline-dependent phospholipase C at 0.1 U m1-1 for 24 h without changing the medium. The background particles were no longer aggregated. Assemblies (circles) again were clearly discernible and randomly distributed. Fig. 8. Thin-section of an astrocytic process. Treatment with choline-dependent phospholipase C at 0.05 U ml -~ for 2 h. The cross-sectioned plasma membrane appeared normal in thin sections (inset enlarged at x 300 000). AC, actin filaments, MT, microtubules, IF, intermediate filaments, M, mitochondria. • 100 000.
464 intermediate filaments, characteristic of astrocytes, were the most centrally located, and were sometimes intermingled with the microtubules (Fig. 8). Choline-dependent phospholipase C from B. cereus also caused astroglial membrane particles to aggregate, but the effective dose (1 U ml 1) was 100 times higher than that of the choline-dependent phospholipase C from C. perfringens. The reason for this difference may be that the two enzymes have different accessibilities to the astroglial membrane because of their different structural configurations. Other phospholipases such as an inositol-dependent phospholipase C, phospholipase D and phospholipase A2 did not cause the astroglial membrane particles to aggregate. Thus, it appears that this astroglial membrane structural change is caused specifically by the action of a choline-dependent phospholipase C. A specialized astrocyte-like cell, the Mfiller cell of the retina, and a transformed astroglioma cell line, C6, were likewise affected by choline-dependent phospholipase C. The Mfiller cells from guinea pig retina, like the C6 glioma cells, did not have assemblies in vitro. However, their membranes showed distinctive particle aggregations and particle-free areas (Figs 9, 11) after 2 h of enzyme treatment. The enzyme was inactive if calcium was omitted from the treatment (Fig. 10). Membrane structure of six other cell types was also assessed at concentrations of up to 1-5 Um1-1 of choline-dependent phospholipase C. A neuron-like cell line, rat PC12 cells (Fig. 12), and rat cerebellar granule neurons were not affected. Other cell types that were not affected include fibroblasts from rat skin and from bovine trachea (Fig. 13), and endothelial cells from bovine brain (Fig. 14) and aorta.
Label-fracture of lectin binding As the gold particles were smaller than assemblies, binding of lectin-gold conjugate to a single assembly could have been recognized. However, the three
TAO-CHENG, BRESSLER and BRIGHTMAN representative lectins used in this study, ConA, WGA and PNA, were not bound to astroglial assemblies. The lectin-gold conjugates were, instead, randomly distributed in the astroglial membranes.
Discussion
Mechanisms which regulate assemblies Two different signal transduction mechanisms that may regulate assembly densities in astrocytes were investigated. These signals were cAMP that activates members of the protein kinase A family and diacylglycerol, which activates members of the protein kinase C family. We found that three drugs which increase cAMP by different mechanisms, forskolin, isoproterenol, and 8-bromo cAMP, increased assembly density and also increased the number of cells expressing assemblies. However, two different reagents that lead to activation of protein kinase C, PMA and phospholipase C, did not alter the assembly densities. Thus it appears that assembly density is regulated by protein kinase A and not by protein kinase C. There are two different mechanisms by which cAMP may act to change assembly density. In one mechanism, assembly precursors are converted to recognizable orthogonal assemblies by a phosphorylation reaction catalyzed by protein kinase A. These precursors or assembly subunits may either reside in the membrane, or in the cytosol. If the precursors are located in the cytosol, a phosphorylation step may be required to translocate the precursors to the membrane or to assemble them into recognizable orthogonal arrays. There are several examples of cytosolic proteins that became modified and attached to plasma membrane (Woodgett et al., 1987). In addition, membrane functions are influenced by phosphorylation. For example, cAMP dependent phosphorylation modifies both beta-adrenergic receptors (Harden,
Figs 9-10. Mfiller cell membranes. Fourth passage, 10 days in vitro, x 100 000. Fig. 9. Treatment with choline-dependent phospholipase C at 0.1 U m1-1 for 2 h. There were aggregates of intramembranous particles and particle-free areas. Fig. 10. The same treatment of choline-dependent phospholipase C except without Ca2+. The membrane particles were not affected. Fig. 11. C6, an astroglial tumor cell line, treated with choline-dependent phospholipase C at 0.1 U ml-~ for 2 h. Membrane particles were aggregated, x 100 000. Fig. 12. PC12, member of a pheochromocytoma cell lines, treated with choline-dependent phospholipase C at 0.1 U ml -I for 2 h. Membrane structure was not affected, x 100 000. Fig. 13. A fibroblast-like cell from a cell line derived from bovine trachea, treated with choline-dependent phospholipase C at 0.5 U m1-1 for 5 h. Membrane structure was not affected, x 100 000. Fig. 14. Brain endothelial (BE) cells, fifth passage, 10 days in culture, treated with choline-dependent phospholipase C at 0.5 U ml -~ for 2 h. Membrane structure was not affected. TJ, tight junction, x 100 000.
466 1983) and K + channels in astrocytes (Neary et aI., 1988). The concentration of assemblies at the site where extracellular K § is siphoned into the retinal glial cell, the Mtiller cell, has been suggested to signify that the assembly constitutes such K § channels (Newman, 1985). In the second mechanism, the phosphorylation induced by protein kinase A increases the production of new assemblies through protein synthesis. The synthesis of other proteins in astrocytes, for example, glutamine synthetase (Juurlink et al., 1981) and GFAP (Pollenz & McCarthy, 1986; Backhovens et al., 1987), are increased after cAMP levels are elevated. We favour this hypothesis of augmented synthesis for two reasons. First, new assemblies require protein synthesis (Anders & Brightman, 1982) and second, assembly density continues to increase in forskolinstimulated astrocytes 24 h after stimulation. Mechanisms dependent on the synthesis of new proteins usually requires hours, while mechanisms dependent on protein phosphorylation usually requires minutes. In the present in vitro experiments, the highest density of assemblies attainable following cAMP stimulation was - 3 0 pum, 2. It is emphasized that this number was only 7.5% that of the highest density (-400 txm 2, Anders & Brightman, 1979) in the most superficial layer of the glia limitans in vivo. Furthermore, astroglial cultures that already had high assembly densities cannot be stimulated by forskolin to produce significantly more assemblies. Thus, it appears that astrocytes, in vitro, have a limited capacity to respond to this signal transduction in the production or intercalation of assemblies. Although the density of assemblies appears to be modulated by cAMP, their distribution is not. When astrocytes were cocultured with brain endothelial cells, aggregates of assemblies with density approaching 400 txm-2 were found (Tao-Cheng et al., 1990). Since similar aggregates were not found in forskolintreated astrocytes, it is not likely that this aggregation was caused by activation of protein kinase A. One possible explanation is that there may be some factor(s) concentrated in the basal lamina, which covers the glia limitans in situ and is also present in patches in astroglial and brain endothelial cocultures. The factors in the basal lamina may hold the assemblies at high densities in the cell membrane directly apposed to basal lamina (Gotow, 1984). Assemblies can be clumped by protein denaturants, but such aggregation is distinct from that brought about by co-culturing astrocytes and brain endothelium. Guanidine and urea cause astroglial assemblies to coalesce into densely packed clumps which appear as large rafts of assemblies with many subunits (Anders & Brightman, 1982), whereas the co-culturing brings about the aggregation of many regular-sized assemblies (Tao-Cheng et al., 1990).
TAO-CHENG, BRESSLER and BRIGHTMAN Non-assembly, background particle aggregation The astroglial, non-assembly, background particles aggregated after choline-dependent phospholipase C treatment. Other phospholipases with different specificities, such as inositol-dependent phospholipase C and phospholipase A2 or D, were inactive even at concentrations that were 10-100 times higher than the choline-dependent phospholipase C. These results suggest the following two conclusions. First, if it is assumed that each enzyme is equally reactive with the astroglial membrane, then the aggregation of background particles in astrocytes is caused by the enzymatic cleaving of a distinct phospholipid, since each enzyme exhibits a restricted specificity. Phosphatidylcholine appears to be the pivotal lipid for integrity of the astroglial plasma membrane since background particles only aggregated after treatment with cholinedependent PLC and not with inositol-dependent PLC. Second, the loss of arachidonic acid (the product of phospholipase A2), from a phospholipid did not induce aggregation of background particles. The loss of arachidonic acid leads to the production of a lysophospholipid, e.g., lysophosphatidylcholine, which, once produced, is not depleted from the membrane because it is quickly reacetylated at the membrane. In contrast, phosphatidylcholine hydrolysis by choline-dependent phospholipase C leads to the depletion of phosphatidylcholine from the membrane, and a longer time is needed for the production of new phosphatidylcholine as replacement. The depletion of this lipid may disrupt the balance of polar and apolar forces at the membrane and allow the background particles to aggregate. At 24 h, the astroglial membrane structure recovered to its normal morphology probably because more phosphatidylcholine was made in that time. Cells of astroglial origin were much more sensitive to phospholipase C treatment than other cell types, including neurons, endothelial cells and fibroblasts. This specific sensitivity may be related to the quantity of phosphatidylcholine or to a different role this phospholipid plays in membrane function in the astrocytes. Phosphatidylcholine may be divided into two pools. One pool is metabolized quickly and has a metabolic function, while the other pool is metabolized more slowly and serves a structural role. It appears that in astrocytes phosphatidylcholine exhibits a more important role in sustaining structure than in other cell types. Conclusion Astroglial assembly production can be augmented by cAMP but not by phorbol esters or diacylglycerols. Thus, protein kinase A but not protein kinase C is involved in assembly production. The assemblies appear to be the product of protein synthesis and not
Forskolin a n d P h o s p h o l i p a s e C affect astroglial m e m b r a n e s p h o s p h o r y l a t i o n since an increase in a s s e m b l i e s w a s still o b s e r v e d 4 8 h after stimulation. In a d d i t i o n , b a c k g r o u n d particles in a s t r o c y t e s a p p e a r u n i q u e l y sensitive to c h a n g e s in p h o s p h a t i d y l c h o l i n e levels. Therefore, this p h o s p h o l i p i d m a y exhibit a different role in a s t r o c y t e s c o m p a r e d to o t h e r cell types.
467
Acknowledgement W e t h a n k Dr K. P e t t i g r e w for advice o n statistical analysis, Dr R. Small for g u i n e a pig M/iller cells, Dr G. G u r o f f for PC12 cells, Dr A. Novelli for rat cerebellar g r a n u l e cell cultures, a n d Dr S. Williams for b o v i n e aortic e n d o t h e l i a l cells.
References ANDERS, J. I. & BRIGHTMAN, M. W. (1979) Assemblies of particles in the cell membranes of developing, mature and reactive astrocytes. Journal of Neurocytology 8, 777-95 ANDERS, J. J. & BRIGHTMAN, M. W. (1982) Particle assemblies in astrocytic plasma membrane are rearranged by various agents in vitro and cold injury in vivo. Journal of Neurocytology 11, 1009-29. BACKHOVENS, H., GHEUENS, J. & SLEGERS, H. (1987) Expression of gial fibrillary acidic protein in rat C6 glioma relates to vimentin and is independent of cell-cell contact. Journal of Neurochemistry 49, 348-54. BIGNAMI, A., ENG, L. F., DAHL, D. & UYEDA, T. (1972) Localization of glial fibrillary acidic protein in astrocytes by immunofluorescence. Brain Research 43, 429-35. BRESSLER, J. P. & TINSELY, P. (1990) Regulation of cAMP levels by protein kinase C in C6 rat glioma cells. Journal of Neuroscience Research 25, 81-6. BRIDGMAN, P. C., KACHAR, B. & REESE, T. S. (1986) The structure of cytoplasm in directly frozen cultured cells. II. Cytoplasmic domains associated with organelle movements., Journal of Cell Biology 102, 1510-21. BRIGHTMAN, M. W. & TAO-CHENG, J.-H. (1990) Mutually imposed structural changes in plasma membranes of astroglia and brain endothelium. In Differentiation and Functions of GliaI Cells (edited by LEVI, G.) pp. 107-15. New York: Alan R. Liss. DERMIETZEL, R. (1974) Junctions in the central nervous system of the cat. III. Gap junctions and membraneassociated orthogonal particle complexes (MOPC) in astrocytic membranes. Cell and Tissue Research 149, 121-35. GERISCH, G. (1987) Cyclic AMP and other signals controlling cell development and differentiation in Dictyostelium. Annual Review of Biochemistry 56, 853-79. GOTOW, T. (1984) Cytochemical characteristics of astrocytic plasma membranes specialized with numerous orthogonal arrays. Journal of Neurocytology 13, 43148. HARDEN, T. K. (1983) Agonist-induced desensitization of the beta-adrenergic receptor-linked adenylate cyclase. Pharmacological Reviews 35, 5-32. JUURLINK, B. H., SCHOUSBOE, A., JORGENSEN, O, & HERTZ, L. (1981) Induction by hydrocortisone of glutamine synthetase in mouse primary astrocyte cultures. Journal of Neurochemistry 36, 13642.
LANDIS, D. M. D. & REESE, T. S. (1974) Arrays of particles in freeze-fractured astrocytic membranes. Journal of Cell Biology 60, 316-20. LESSEPS, R. J. (1967) The removal of phospholipase C of a layer of lanthanum-staining material external to the cell membrane in embryonic chick cells. Journal of Cell Biology 34, 173-83. MCCARTHY, M. D. & DE VELLIS, J. (1980) Preparation of separate astroglial and oligodendroglial cell cultures from rat cerebral tissue. Journal of Cell Biology 85, 890-902. NEARY, J. T., NORENBERG, L. O. B. & NORENBERG, M. D. (1988) Protein kinase C in primary astrocyte cultures: cytoplasmic localization and translocation by a phorbol ester. Journal of Neurochemistry 50, 1179-84. NEWMAN, A. E. (1985) Membrane physiology of retinal glial (Mtiller) cells. Journal of Neuroscience 5, 2225-39. PINTO DA SILVA, P. & KAN, F. W. K. (1984) Label-fracture: a method for high resolution labeling of cell surfaces. Journal of Cell Biology 99, 1156-61. POLLENZ, R. S. & MCCARTHY, K. D. (1986)Analysis of cyclic AMP-dependent changes in intermediate filament protein phosphorylation and cell morphology in cultured astroglia. Journal of Neurochemistry 47, 9-17. SMALL, R. K., PATEL, P. & WATKINS, B. A. (1991) Response of Miiller cells to growth factors alters with time in cultme. Glia 4, 469-83. TAO-CHENG, J,-H., BRESSLER, I. P. & BRIGHTMAN, M. W. (1987a) Structural changes of astroglia membranes in vitro. Society for Neuroscience Abstracts 13, 120. TAO-CHENG, J.-H., NAGY, Z. & BRIGHTMAN, M. W. (1987b) Tight junctions of brain endothelium in vitro are enhanced by astrocytes. Journal of Neuroscience 7, 3293-9. TAO-CHENG, J.-H., BRESSLER, J. P. & BRIGHTMAN, M. W. (1988) Characterization of astroglial membrane assemblies by agents which activate protein kinase C. Society for Neuroscience Abstracts 14, 430. TAO-CHENG, J.-H., NAGY, Z. & BRIGHTMAN, M. W. (1990) Astrocytic orthogonal arrays of intramembranous particle assemblies are modulated by brain endothelial cells in vitro. Journal of Neurocytology 19, 143-55. WOODGETT, J. R., HUNTER, T. & GOULD, K. L. (1987) Protein kinase C and its role in cell growth. In Cell Membranes Methods and Reviews Vol. 3 (edited by ELSON, E., FRAZIER, W. & GLASER, L.) pp. 215--340. New York: Plenum Press.
