Anat Embryol (1992) 185 : 355-361 Anatomy and Embryology �9 Springer-Verlag 1992

Parvalbumin immunoreactive neurons and fibres in the teleost cerebellum

J.R. Alonso, R. Ar~valo, J.G. Brifi6n, J. Lara, E. Weruaga, and J. Aij6n

Department of Cell Biology, University of Salamanca, E-37007 Salamanca, Spain

Accepted July 11, 1991

Summary. The distribution of parvalbumin- (PV) im- munopositive cell bodies and fibres in the cerebellum of two species of freshwater teleosts (Salmo gairdneri and Barbus meridonalis) was studied using a monoclonal antibody and the avidin-biotin immunoperoxidase tech- nique. A clear laminated pattern of PV immunoreactivi- ty was observed. After PV-immunostaining, Purkinje cells were strongly labelled in their cell bodies, the initial segments of the axons and the dendritic trees. In the molecular layer, only the dendritic branches of the Pur- kinje cells were PV-positive. In the granule cell layer, extensive axonal plexuses and scattered cell bodies were observed. Most of the immunopositive perikarya were unequivocally identified as displaced Purkinje cells, whereas a reduced number of smaller neurons with un- stained dendrites was also found. Eurydendroid cells, the efferent neurons of the teleost cerebellum, were nega- tive; however, they were impinged upon by numerous PV-positive boutons, corresponding to terminals of Pur- kinje cell axons. Parallel fibres and climbing fibres, as well as stellate cells and granule cells were negative. Bas- ket cells (or deep stellate cells) whose existence in the teleost cerebellum is discussed, were also not observed. The immunoreactivity distribution pattern for PV in the teleost cerebellum differs from previous observations on the localization of this protein in the cerebellum of am- niotes.

[

Key words: Parvalbumin Cerebellum - Purkinje cell Calcium-binding Teleost

Introduction

The effect of calcium ions is exerted through membrane and cytosolic targets, including in this latter group pro- tein kinase C, calpains, calmodulin, and other calcium-

Offprint requests to: J. Aijdn

binding proteins (Kennedy 1989). Some calcium-modu- lated proteins such as calmodulin are ubiquitous, where- as others, such as calbindin D-28k, calretinin or parval- bumin (PV) are only present in definite locations. PV is distributed in several regions of the mammalian CNS (Cello 1990). In these different locations, PV can be used as a specific marker for several neuronal populations since it is only present in distinct cell types, or even in a partial subpopulation of a neuronal type (Celio and Heizmann 1981; Heizmann 1984; Sloviter 1989; Alonso et al. 1990; Celio 1990).

The function of this calcium-binding protein remains to be elucidated. One possible approach for knowing the exact role of PV in the CNS is a systematic mapping of positive cells, and a search for a correlation with known functional mechanisms (Zuschratter et al. 1985). In this sense, a broadening of the available studies, in- cluding different phylogenetic groups, is necessary. However, although there are biochemical data indicating the presence of calcium-binding proteins in the teleos- tean brain (Gosselin-Rey et al. 1978), there are very few immunohistochemical studies on the identification of positive cells in the brain of anamniotes (Maler et al. 1984; Goto et al. 1988; Denizot et al. 1988; Rodriguez- Moldes etal. 1990a, b).

In mammals, one of the regions with highest PV im- munoreactivity is the cerebellum, Which shows PV im- munopositivity in several neuronal types, including bas- ket, stellate and Purkinje cells (Celio and Heizmann 1981; Heizmann 1984). Although the basic structure of the cerebellum is easily comparable in all vertebrates, that of fish shows the greatest range of structural varia- tion found in any class (Schnitzlein and Faucette 1969; Lee and Bullock 1984). Furthermore, certain fundamen- tal differences at both anatomical and physiological lev- els have been described between the cerebellum of te- leosts and that of amniotes, specially mammals (Nieuw- enhuys etal. 1974; McGlone and Paul 1980; Finger 1983; Lee and Bullock 1984; Uchiyama etal. 1988; Meek and Nieuwenhuys 1991). These differences agree

356

well with previous data suggesting phylogenetic varia- tions in the distribution pattern of neuroactive sub- stances in the cerebellum (Schulman et al. 1981).

The aim of the present work was to examine the dis- tribution of PV in the cerebellum of the rainbow trout (Salmo gairdneri) and the mediterranean barbel (Barbus meridionalis), and to compare the results with previous data reported in other vertebrates. Our conclusions may provide further information on the localization and function of this calcium-binding protein within the CNS of anamniotes.

Materials and methods

Animals and tissue preparation. Twelve adult rainbow trouts (Salmo gairdneri Richardson) with 210-245 g body weight, obtained from commercial sources (Fisheries "Sieteigtesias", Salamanca), and ten mediterranean barbels (Barbus meridionalis Risso) with 295-480 g body weight, captured in the river Tormes (Salamanca), were used for the present study. They were kept under standard laboratory conditions (12/12 h light/dark cycle and fed with trout chow). The animals were deeply anaesthetized by immersion in a 0.03% solu- tion of tricaine methanesulfonate (MS-222, Sandoz) and perfused transcardially as described elsewhere (Alonso et al. 1989). A fixa- tive containing 4% paraformaldehyde, 0.08% glutaraldehyde, and 15% saturated picric acid in 0.12 M phosphate buffer (pH 7.2) (Somogyi and Takagi 1982) was used in eight animals from each species. After perfusion, the brains were removed, the cerebellum dissected out and stored in glutaraldehyde-free fixative for two additional hours. Finally, the blocks were rinsed several times in phosphate buffer. With a vibratome (Campden Instruments) or a cryostat (Bright), 40-gm sections were cut in sagittal and trans- verse planes, thoroughly washed in phosphate buffer, and stored overnight at 4 ~ C.

The remaining animals were perfused with 4% paraformalde- hyde in 0.1 M phosphate buffer (pH 7.3). Their brains were extract- ed and embedded in paraffin, sectioned at 9 gm, and mounted on slides. The hydrated sections were processed in a humid chamber in the same way as the free-floating vibratome and cryo- stat sections. Similar results were obtained for all the sections.

Antibodies and immunoeytochemical procedure. Free-floating sec- tions were incubated with a monoclonal mouse anti-carp muscle parvalbumin (McAB 235). This antibody has been fully character- ized in recent papers (Celio et al. 1988; Tinner et al. 1990). The PV-antibody was diluted at 1:5000 or 1:10000 in 0.1 M phosphate buffer containing 10 % normal horse serum, incubating the sections for 48 h at 4 ~ C. The sections were then washed in phosphate buffer and processed according to the avidin-biotin immunoperoxidase (ABC) method (Hsu et al. 1981). Firstly, the sections were incubat- ed with biotinylated anti-mouse immuno-gammaglobulin (Vector Laboratories, Burlingame, U.S.A.) diluted 1:250 for 3 h at 20 ~ C, and Vectastain ABC reagent (1 : 250) for two additional hours. Tis- sue-bound peroxidase was visualized by incubating the sections with 0.07% 3,3' diaminobenzidine and 0.003 % hydrogen peroxide in Tris buffer (0.1 M, pH 7.6) for 10 min. Finally, the sections were dehydrated in increasing ethanol series and mounted with Araldite between two plastic foils, or with Entellan using coverslips.

The specificity of the immunostaining was controlled by omit- ting the anti-PV antibody or the second antibody in the incubation bath. In both cases, no residual immunoreactivity was found. Moreover, possible interference by endogenous peroxidases was ruled out by staining some sections beginning with the diaminoben- zidine step. No reaction was visualized. Finally, vibratome sections of different regions of the rat brain such as the cerebellum, olfacto- ry bulb, hippocampal and septal regions, were cut and processed for immunostaining together with the sections of the fish cerebel-

lum, since the former regions contain characteristic patterns of PV immunoreactivity (Celio and Heizmann 1981 ; Endo et al. 1986; Kosaka et al. 1987; Katsumaru etal. 1988; Celio 1990; Alonso et al. 1990); they were therefore used as additional controls of the immunostaining procedure.

Neuronal size was measured by means of a Zeiss ocular mi- crometer.

Results

The following layers were differentiated in the teleost cerebellum: Molecular layer: mainly built up of dendrites of Purkinje cells, eurydendroid cells, climbing fibres and parallel fibres. In addition, at least one neuronal population, the stellate ceils, is also located in this stratum. The presence of basket cells is discussed (see Finger 1983 and Meek and Nieuwenhuys 1991). Purkinje cell layer or ganglionic layer: containing large neurons - mostly Purkinje cells and eurydendroid cells - and perhaps a population of smaller neurons, classified as deep stellate cells or basket neurons (see above). Granule cell layer: showing numerous small neurons, the granule cells and scattered larger neurons, identified as Golgi Type II neurons.

Since the labellings for PV are very similar in both species studied, they are described together.

ParvaIbumin imrnunoreactivity

With the PV antibody employed, all Purkinje cells were strongly immunostained in their cell bodies, dendritic arborizations and the initial segments of the axons (Figs. 1, 2). These neurons were irregularly arranged in two to four levels in the ganglionic layer (Figs. 1 A, 2A, B, D), although with regional differences in the thickness and development of this stratum. Thus, in the corpus (Figs. 1A, 2A) Purkinje cell bodies and primary den- drites were very regularly oriented, whereas in the valvu- la and lobus, Purkinje cells were not so densely packed, their dendrites course in oblique directions, and the cell bodies were disposed at several levels, including dis- placed elements in the molecular layer (Fig. 1 E).

The positive Purkinje cells showed normally round, pyriform or fusiform somata (20-35 ~tm maximum di- ameter), from which arose one thick dendrite. These lat- ter prolongations extended mostly perpendicular to the surface of the cerebellum, forming a compact palisade- like structure perpendicularly oriented to the parallel fibre system (Fig. 2A). The complete dendritic arboriza- tion was PV-immunostained; however, immunostained dendritic spines were not observed. The initial segment of the axon was also labeled arising from the basal pole of the cell body (Figs. 1 B, D, 2A, B, D) or, rarely, from a thick primary dendrite. This PV-immunostained initial prolongation of the axon extended primarily towards the granule cell layer, but it was also found turning to- wards the surface and coursing horizontally in the gan- glionic layer for a distance. In addition, collaterals were also observed arising from the immunostained axons.

357

Fig. 1. A Photomontage of a transverse section of the rainbow trout brain showing PV-immunoreactivity in the teleost cerebellum. g, granule cell layer; m, molecular layer; p, Purkinje cell layer; v, fourth ventricle, Arrows point to PV-immunostained fascicles. Open arrow: displaced Purkinje cell. • 36.5. B Displaced Purkinje cells in the granule cell layer, p, Purkinje cell layer, x 175. C Im- munopositive small neuronal cell bodies (arrows) in the granule

cell layer, x 125. D Stained neuron (open arrow) with a very long dendrite (small arrows) parallel to the Purkinje cell layer, x 90. E PV-positive neurons (Purkinje cells) in the Purkinje cell layer (arrows) and in the molecular layer (open arrows) in the valvula cerebelli, x 90. F Positive Purkinje cell with two primary dendrites (arrows). x 175

358

Thus, the Purkinje cell axons could be followed for long distances, most of them coursing horizontally in the out- er (subganglionic) region of the granule cell layer (Fig. 2 D). However, there were also regional differences between the density and the disposition of the immunos- tained axons (compare Figs. 2A-D).

Some PV-positive neurons were observed in the gan- glionic layer differing in several aspects from the classical description of the teleostean Purkinje cell. Thus, we have seen Purkinje cells with two primary dendrites arising from the cell body (Fig. 1 F), or with one primary den- drite showing a large diameter and extending, horizon- tally or vertically, for more than 50 gm without branch- ing (Fig. 1D). The primary dendrites of these cells branched in the ganglionic layer or in the deeper region of the molecular layer, forming a flattened dendritic tree. Their terminal prolongations also extended perpendicu- lar to the cerebellar surface, but not so regularly and compact as in other Purkinje cells. All these characteris- tics were previously considered typical for eurydendroid cells, demonstrating a higher variability of the Purkinje cells.

We differentiated a second population of large neu- rons in the teleost ganglionic layer, which were PV-nega- tive and were identified according to their size, shape and location as eurydendroid cells. These PV-immunon- egative neurons were less numerous than the Purkinje cells, showed fusiform, round, pyriform or triangular cell bodies with their maximum diameter oriented paral- lel to the boundaries between the ganglionic layer and the adjacent strata. Eurydendroid cells were mostly found in the Purkinje cell layer (Fig. 2 E), but they were also observed in the granule cell layer (Fig. 2F). The main characteristic of these neurons was the consider- able number of PV-positive boutons impinging onto the unstained neuronal bodies (Fig. 1 E and 1 F).

PV-immunolabeled elements were also observed out- side the ganglionic layer. Thus, in the molecular layer, it was possible to follow the dendritic ramifications of the Purkinje neurons until the proximity of the glia limi- tans. However, neurons intrinsic to this stratum such as stellate cells, were not immunostained. In the granule cell layer, abundant PV-positive axonal plexuses were observed, specially in the outer zone (Figs. 1 A, 2). In addition, labelled perikarya were also found. Most of them were unequivocally identified as displaced Purkinje cells by their size and morphological characteristics (Fig. 1 B), while other cells displayed smaller cell bodies (but considerably larger than the unlabelled granule cells) an their distal dendrites were not stained (Fig. 1 C).

Fig. 2. A Section of the corpus cerebelli showing positive PV-label- ling in the molecular (m), Purkinje cell (p) and granule cell (g) layers, x 135. B Purkinje cells more irregularly disposed (arrow) and positive axons paralM to the Purkinje cell layer (open arrows). x 175. C Network of parvalbuminergic fibres in the inner granule cell layer (arrow). • 175. D Dense groups of positive axons (arrow) just under the Purkinje cell layer, x 250. E Negative neuron (ar- row), identified as a eurydendroid cell, in the Purkinje cell layer. x 350. F Two eurydendroid cells (arrows) surrounded by parvalbu-

min-positive boutons in the granule cell layer. • 350

359

Discussion

We have performed an immunohistochemical mapping of antigenic site for PV in the teleost cerebellum and have found them to be present in distinct cell popula- tions. The distribution pattern of immunoreactivity differs partially from that of this protein in amniotes, as well as from other calcium-binding proteins such as calmodulin, S-100 protein and calcineurin (Goto et al. 1988).

Although the basic structure of the cerebellum in lower and higher vertebrates is easily comparable, show- ing several similarities, differences have been observed in the morphology, electrophysiology and neuronal or- ganization of this nervous centre (Finger 1983). These include the presence of the valvula cerebelli and the pre- cerebellar nucleus lateralis valvulae, structures absent from other vertebrates (Meek and Nieuwenhuys 1991), of a distinct cytoarchitecture, showing new neuronal types such as the eurydendroid cell, and the absence of an aggregation of efferent neurons that could be de- scribed as a nucleus (Nieuwenhuys et al. 1974). Electro- physiological experiments have also demonstrated pecu- liarities in the circuitry of the teleostean cerebellum. Ex- cept for the valvula, where inputs from all sensory mo- dalities converge, different sensory afferences terminate in different parts of the teleost cerebellum with only a little overlapping. In the mammalian cerebellum on the other hand several studies, using field potentials or unitary responses, have shown the interaction in the cer- ebellum of inputs from different sensory modalities (Lee and Bullock 1984). The way in which these different sensory inputs are integrated in the fish cerebellum is unknown. Thus, observations on the cerebellum of am- niotes cannot be directly extrapolated to the teleost cere- bellum.

A comparison of our observations in Salmo gairdneri and Barbus meridionalis with previous data on the immu- noreactivity of PV in the cerebellum of higher verte- brates, points to both similarities and differences. All the Purkinje cells were positively labeled for PV in mam- mals and birds, as we have reported in the rainbow trout and the mediterranean barbel. In addition, biochemical studies have demonstrated that these neurons show the highest activity of this calcium-binding protein in the rat (Celio and Heizmann 1981) and in birds (Braun et al. 1985). PV seems to be present in the Purkinje cells of all vertebrates, although due to the presence of euryden- droid cells the basic cerebellar circuity is different in the teleosts and the teleostean Purkinje cells can be clas- sified as interneurons.

Furthermore, one type of neuron peculiar to the te- leost cerebellum, the eurydendroid cell, is not immunos- tained for PV in our material. This is important, since until it had been studied with Golgi techniques (Nieuw- enhuys et al. 1974) this kind of cell was not differentiated from the Purkinje cell. Several differences have been de- scribed between the Purkinje cells and the eurydendroid cells, but some morphological data, such as dendritic arborization arising from a single primary dendrite (Pur- kinje cells) or from two or more (eurydendroid cells)

360

(Nieuwenhuys et al. 1974) are not good indicators, as shown in our results. Additionally, axonal labelling of cerebellar neurons has shown that few if any Purkinje cell axons leave the cerebellum, whereas the euryden- droid cells provide the efferent fibres of this structure (Finger 1983). Using PV-immunocytochemistry it is pos- sible to follow the stained axons of Purkinje cells for long distances and to see their endings on the unstained eurydendroid cells. On the other hand, the physiology of Purkinje cells and eurydendroid cells is different: to- gether with granule cells (PV-negative in all studied spe- cies) eurydendroid cells are the only cerebellar neurons capable of producing excitatory postsynaptic potentials (Nieuwenhuys et al, 1974; Finger 1983). Finally, the eur- ydendroid cells, in spite of their position, have been com- pared according to their significance in the cerebellar integration to the deep cerebellar nuclei neurons of other vertebrates (Nieuwenhuys and Nicholson 1969). In this respect, it is interesting to note that no PV-positive neu- ron can be observed in the medial, anterior and posterior interposed and dentate nuclei; and, on the other hand, neurons in these nuclei have a large number ofimmunos- tained terminals impinging on their perikarya (Celio 1990). The available data are insufficient to substantiate a clear homology of this neuronal type with other cells of amniotes and further studies, also involving immuno- cytochemistry, would be necessary. For this purpose, PV seems to be an excellent marker to differentiate, at the ultrastructural level, cell bodies, dendritic trees, axonal prolongations, and synaptic contacts belonging to Pur- kinje cells or to eurydendroid cells which are densely intermingled.

With regard to other PV-positive cerebellar neurons, in the mammalian molecular layer, basket cells and stel- late cells were labelled for PV (Heizmann 1984; Endo et al. 1986) whereas in our PV-immunostained material no labelled perikarya, apart from displaced Purkinje cells, were observed in this stratum. Additionally, the granule cells diplayed no immunolabelling in any of the studied phylogenetic groups (Celio and Heizmann 1981). The absence of stained elements which can be identified as basket cells confirms that they are not present in the teleost cerebellum, as suggested by other techniques (Meek and Nieuwenhuys 1991). On the other hand, it is unknown whether stellate cells project both on Pur- kinje dendrites and eurydendroid dendrites, or only on one of these elements (Meek and Nieuwenhuys 1991). Thus, it is not clear if these neurons can be considered a homogeneous group in all vertebrates or, as our results suggest, there are significant phylogenetic differences.

Calcineurin has been also observed in the cerebellum of other classes of vertebrates, including fish (Goto et al. 1988). Calbindin D-28k (CaBP-D28K), another calcium- binding protein which has been found to have a comple- mentary distribution to PV in different brain regions (Celio 1989, 1990), has been demonstrated in fish, am- phibia, reptiles, birds and mammals (Parmentier et al. 1987; Brauth etal . 1988; Denizot etal . 1988; Cello 1990; Rodrlguez-Moldes et al. 1990a, b). In the cerebel- lum, it has been detected in the Purkinje cells of amphi- bia, birds and mammals (Jande et al. 1981 ; Baimbridge

and Miller 1982; Baimbridge et al. 1982; Legrand et al. 1983; Garcia-Segura et al. 1984; Zuschratter et al. 1985). However, a recent work (Rodriguez-Moldes et al. 1990 a) indicated that in cartilaginous fish, on the contrary, Pur- kinje cells and all other cerebellar neurons are CaBP D-28k-negative. Other phylogenetic differences have been reported for the distribution of this calcium-bind- ing protein, since in the chicken (Garcia-Segura et al. 1984) and frog (Legrand et al. 1983) there are no CaBP D-28k-positive Golgi cells, whereas in the rat and hu- man, a population of Golgi cells was positively immun- ostained (Garcia-Segura et al. 1984). We have tested in teleosts the monoclonal antibody (McAB 300) (see Celio 1990 and Cetio et al. 1990 for characterization of this antibody) against CaBP D-28k (very much used in high- er vertebrates, and highly specific since it does not cross- react with the homologous protein calretinin) and found that it stained selectively glial populations along the whole fish brain, including the cerebellum. This suggest that contrary to the PV-antibody (generated against a teleostean protein), the anti-CaBP antibody (anti-chick gut calbindin D-28k) recognizes in teleosts a different epitope, or a different substance, to that recognized in mammals. Thus, biochemical studies seem to be needed in order to clarify the homologies and differences of the primary sequence of this calcium-binding protein in the different vertebrate classes.

PV, CaBP D-28k, S-100 protein, calcineurin, calmo- dulin and other low molecular weight calcium-binding proteins are biochemically related due to their common evolutionary origin (Goodman et al. 1979); however, their cellular and subcellular distributions are quite dif- ferent, indicating different physiological functions (Heiz- mann 1984), These characteristics, and the presence of cells containing at least three different calcium-binding proteins, suggest complex interactions between the dif- ferent molecules, directly related to the physiology of each cell population.

Acknowledgements. The authors are greatly indebted to Prof. M.R. Celio (University of Fribourg, Switzerland) for kindly providing the antibodies against parvalbumin and calbindin D-28k. This work was supported by the Universidad de Salamanca (grants to J.R. Alonso and J. Aij6n) and the Spanish DGICYT (PM88-0154).

References

Alonso JR, Covefias R, Lara J, De Le6n M, Aij6n J (1989) Distri- bution of VIP-like immunoreactivity in the olfactory bulb of the rainbow trout. Brain Res 490:385-389

Alonso JR, Covefias R, Lara J, Aij6n J (1990) Distribution of parvalbumin immunoreactivity in the rat septal area. Brain Res Bull 24: 41-48

Baimbridge KG, Miller JJ (1982) Immunohistochemical localiza- tion of calcium-binding protein in the cerebellum, hippocampal formation and olfactory bulb of the rat. Brain Res 245 : 223-229

Baimbridge KG, Miller J J, Parkes CO (1982) Calcium-binding pro- tein distribution in the rat brain. Brain Res 239:519-525

Braun K, Scheich H, Schachner M, Heizmann CW (1985) Distribu- tion of parvalbumin, cytochrome oxidase activity and [14C]-2- deoxyglucose uptake in the brain of the zebra finch. Cell Tissue Res 240:101-115

Brauth SE, Kitt CA, Gerfen CR (1982) Calcium binding protein on the basal ganglia system of a non-mammalian vertebrate:

361

an immunohistochemical study in the reptile Caiman crocodilus. Brain Res 452:367-372

Celio MR (1989) Calcium binding proteins in the brain. Arch Ital Anat Embryol 94:227-236

Cello MR (1990) Calbindin D-28k and parvalbumin in the rat nervous system. Neuroscience 35:375~475

Celio MR, Heizmann CW (1981) Calcium-binding protein parval- bumin as a neuronal marker. Nature (London) 293 : 300-302

Celio MR, Baier W, Sch/irer L, De Viragh PA, Gerday Ch (1988) Monoclonal antibodies directed against the calcium binding protein parvalbumin. Cell Calcium 9 : 81-86

Celio MR, Baier W, SehS, rer L, Gregersen H J, De Viragh PA, Norman AW (1990) Monoclonal antibodies directed against the calcium binding protein Calbindin D-28k. Cell Calcium 11 : 599-602

D~nizot JP, Bratton BO, Br~hier A, Thomasset M (1988) Immuno- histochemical demonstration of calbindin-D 28K (CABP28K) in the spinal cord motoneurons of teleost fish. Cell Tissue Res 254:629 634

Endo T, Kobayashi M, Kobayashi S, Onaya T (1986) Immunocyto- chemical and biochemical localization of parvalbumin in the retina. Cell Tissue Res 243:213-217

Finger TE (1983) Organization of the teleost cerebellum. In: North- cutt RJ, Davis RE (eds) Fish Neurobiology, vol. 1 : Brain stem and sense organs. The University of Michigan Press, Ann Ar- bor, pp 261 284

Garcia-Segura LM, Baetens D, Roth J, Norman AW, Orci L (1984) Immunohistochemical mapping of calcium-binding protein im- munoreactivity in the rat central nervous system. Brain Res 296 : 75-86

Goodman M, P6ch+re JF, Haiech J, Demaille JG (1979) Evolution- ary diversification of structure and function in the family of intracellular calcium-binding proteins. J Mol Evol 13 : 331-352

Gosselin-Rey C, Piront A, Gerday C (1978) Polymorphism of par- valbumins and tissue distribution. Characterization of compo- nent I. isolated from red muscles of Cyprinus carpio L. Biochim Biophys Acta 532 : 293 304

Goto S, Matsukado Y, Uemura S, Mihara Y, Inoue N, Ikeda J, Miyamoto E (1988) A comparative immunohistochemical study of calcineurin and S-100 protein in mammalian and avian brains. Exp Brain Res 69:645-650

Heizmann CW (1984) Parvalbumin, an intracellular calcium-bind- ing protein; distribution, properties and possible roles in mam- malian cells. Experientia 40 : 910-921

Hsu SM, Raine L, Fanger H (1981) The use of avidin-biotin-peroxi- dase complex (ABC) in immunoperoxidase techniques: A com- parison between ABC and unlabeled antibody (peroxidase) pro- cedures. J Histochem Cytochem 29:57%590

Jande SS, Tolnai S, Lawson DEM (1981) Immunohistochemical localization of vitamin D-dependent calcium-binding protein in duodenum, kidney, uterus and cerebellum of chickens. Histo- chemistry 71:99-116

Katsumaru H, Kosaka T, Heizmann CW, Hama K (1988) Immun- ocytochemical study of GABAergic neurons containing the cal- cium-binding protein parvalbumin in the rat hippocampus. Exp Brain Res 72:347-362

Kennedy MB (1989) Regulation of neuronal function by calcium. Trends Neurosci 12(11):417-424

Kosaka T, Katsumaru H, Hama K, Wu JY, Heizrnann CW (1987) GABAergic neurons containing the Ca 2 +-binding protein par- valbumin in the rat hippocampus and dentate gyrus. Brain Res 419:119-130

Lee LT, Bullock TH (1984) Sensory representation in the cerebel- lum of the catfish. Neuroscience 13:157-169

Legrand C, Thomasset M, Parkes CO, Clavel MA, Rabi6 A (1983) Calcium-binding protein in the developing rat cerebellum. Cell Tissue Res 233:389-402

Maler L, Jande S, Lawson EM (1984) Localization of vitamin D-dependent calcium-binding protein in the electrosensory and electromotor system of high frequency gymnotid fish. Brain Res 301:166-170

McGlone FP, Paul DH (1980) Morphology and electrophysiology of the cerebellum of the perch, Percafluviatilis. J Physiol (Lon- don) 303 : 23

Meek J, Nieuwenhuys R (1991) Palisade pattern of mormyrid Pur- kinje cells: A correlated light and electron microscopic study. J Comp Neurol 306:156-192

Nieuwenhuys R, Nicholson C (1969) Aspects of the histology of the cerebellum of mormyrid fishes. In : Llin~is R (ed) Neurobio- logy of cerebellar evolution and development. Am Med Assoc, Chicago, pp 135-169

Nieuwenhuys R, Pouwels E, Smulders-Kersten E (1974) The neuro- nal organization of cerebellar lobe C1 in the mormyrid fish Gnathonemus petersii (Teleostei). Z Anat Entwicklungsgesch 144:315-336

Parmentier M, Ghysens M, Rypens F, Lawson DEM, Pasteels JL, Pochet R (1987) Calbindin in vertebrate classes: immuno- histochemical localization and Western blot analysis. Gen Comp Endocrinol 65 : 399-407

Rodriguez-Moldes I, Timmermans JP, Adriaensen D, De Groodt- Lasseel MHA (1990a) Immunohistochemical localization of calbindin-D28k in the brain of a cartilaginous fish, the dogfish (Scyliorhinus canicula L.). Acta Anat 137:293-302

Rodriguez-Moldes I, Timmermans JP, Adriaensen D, De Groodt- Lasseel MHA, Scheuermann DW, Anadon R (1990b) Asym- metric distribution of calbindin-D28k in the ganglia habenulae of an elasmobranch fish. Anat Embryol 181:389-391

Schnitzlein HC, Faucette JR (1969) General morphology of the fish cerebellum. In: Llin/ts R (ed) Neurobiology of Cerebellar Evolution and Development. Am Med Ass, Chicago, pp 77-106

Schulman JA, Finger TE, Brecha NC, Karten HJ (1981) Enkepha- lin imrnunoreactivity in Golgi cells and mossy fibres of mamma- lian, avian, amphibian and teleost cerebellum. Neuroscience 6: 2407-2416

Sloviter RS (1989) Calcium-binding protein (calbindin-D28k) and parvalbumin immunocytochemistry: localization in the rat hip- pocampus with specific reference to the selective vulnerability of hippocampal neurons to seizure activity. J Comp Neurol 280:183-196

Somogyi P, Takagi H (1982) A note on the use of picric acid- paraformaldehyde-glutaraldehyde fixative for correlated light and electron microscopic immunocytochemistry. Neuroscience 7:1779-1784

Tinner R, Oertle M, Heizmann CW, Bosshard HR (1990) Ca 2+- binding site of carp parvalbumin recognized by monoclonal antibody. Cell Calcium 11:19-23

Uchiyama H, Matsutani S, Ito H (1988) Pretectum and accessory optic system in the filefish Navodon modestus (Balistidae, Te- leostei) with special reference to visual projections to the cere- bellum and oculomotor nuclei. Brain Behav Evol 31 : 170-180

Zuschratter W, Scheich H, Heizmann CW (1985) Ultrastructural localization of the calcium-bindig protein parvalbumin in neu- rons of the song system of the zebra finch, Poephila guttata. Cell Tissue Res 241:77-83