THE JOURNAL OF COMPAR.ATIW3 NEUROLOGY 302:473-484 (1990)

Morphology and Distribution of the GlossopharpgealNerve Merent andEfferentNeurons in the Mexican Salamander,Axolotl: A Cobaltic-LysineStudy TAKATOSHI NAGAI AND TOSHIYA MATSUSHIMA Department of Physiology, Teikyo University School of Medicine, Tokyo 173 (T.N.), and Zoological Institute, Faculty of Science, University of Tokyo, Tokyo 113 (T.M.),Japan

ABSTRACT Cobaltic-lysine complex was used to label the afferent and efferent components of the glossopharyngeal nerve in the ganglion and brainstem of the Mexican salamander, axolotl (Ambystomamexicanurn).The distribution of afferent cell bodies in the combined glossopharyngeal-vagus ganglion (the IX-X ganglion) was reconstructed from serial sections, and the sizes of the cell bodies were measured. The central projection of afferents and the location of efferent cell bodies were determined by the tracer. The afferent cell bodies in the ganglion were medium-sized (ca. 25 km). Cell bodies with a single process were seen. The ganglion was not clearly divided into superior and inferior ganglia, as is observed in mammals and frogs, but comprised a single ganglion. Labelled cells were diffusely distributed in the rostra1 part of the IX-X ganglion. A few labelled cells also were seen in the caudal part, where the vagus nerve fibers and cell bodies were mainly distributed. Double labellings of the glossopharyngeal and vagus nerves with HRP and cobaltic-lysine demonstrated that the ganglion cells of each nerve are not clearly separated in the IX-X ganglion. In the brainstem, the majority of afferent fibers formed thick ascending and descending limbs in the solitary fasciculus. The remaining afferent fibers formed a thin bundle in the spinal tract of the trigeminal nerve, which had a short ascending limb and a long descending limb. These two bundles had terminal areas in the ipsilateral brainstem: in the dorsal gray matter for the solitary fasciculus and in the lateral funiculus for the spinal tract of the trigeminal nerve, respectively. The cell bodies of the efferent neurons possessed developed dendritic arborizations in the ventrolateral white matter, and formed a longitudinal cell column in the ventrolateral margin of the gray matter. Thus, the glossopharyngeal nerve system in the axolotl assumes a primordial form in its ganglions, but its topographical organization in the brainstem is basically similar to that in anurans. Key words: taste, Ambystoma, brainstem, neuroanatomy, chemoreceptors

For the experiments to elucidate neural mechanisms of taste in vertebrates, a variety of animals in the classes Mammalia, Amphibia, and Pisces are being used. In amphibians, anurans provided a considerable amount of knowledge, in particular, on the transduction mechanism in taste receptor cells (Sato and Beidler, '75; Akaike et al., '76; Sato, '76; Avenet et al., '88). Urodeles, another major group of amphibians, have also been used in many physiological experiments (Kuffler, et al., '66; Werblin and Dowling, '69; Diamond et al., '76; Dennis and Yip, '78; Getchell and Shepherd, '78; Model, '78; Attwell et al., '87; Kauer, '881, because the nerve cells and sensory receptor cells are

o 1990 WILEY-LISS, INC.

usually large, However, it is relatively recently that the urodeles were introduced into the electrophysiological study of taste transduction (West and Bernard, '78; Roper, '83; Kinnamon and Roper, '87, '88; Sugimoto and Teeter, '87). In the urodeles, taste receptor cells are also large, and form large taste buds of a simpler structure (Farbman and Yonkers, '71; Toyoshima et al., '87). These features make Accepted Aug 28,1990. Address reprint request to Dr. Takatoshi Nagai, Dept. of Physiology, Teikyo University School of Medicine, Kaga 2-11-1, Itabashi-ku, Tokyo 173, Japan.

T. NAGAI AND T. MATSUSHIMA

474

the cells accessible to various experimental manipulations, i.e., microelectrode impalements, chemical stimulation, or histological study, and thus have promoted the recent advances in the cell biology of taste receptors (Roper, '89). A regenerative spike was first demonstrated in the taste receptor cell of the mudpuppy, Necturus (Roper, '83).Such an excitatory property of taste cells had not been reported previously in other preparations (in rat, Ozeki, '71; in frog, Sato and Beidler, '75; and Akaike et al., '76). Contradictory results in these animals arose because the small taste cells in rats and frogs may have been damaged by impalement with the microelectrode and lost their excitability, but not the larger cells in the mudpuppy (Roper, '83). The large taste cells in urodeles also favor morphological studies. Fluorescent dye was successfully injected to show a functional coupling between the taste receptor cells (Yang and Roper, '87). The fine structure of the taste cells and their synapses was also studied with the electron microscope (Farbman and Yonkers, '71; Toyoshima et al., '87; Delay and Roper, '88). Sensory information transduced in the taste cells is transferred to the brain by the primary gustatory afferents. The gustatory neurons in the peripheral as well as in the central nervous system have not been studied with intracellular recordings, because these neurons are generally small, and hence not amenable to electrode impalement. This drawback of gustatory neurons in general has limited our knowledge of their cellular properties. To investigate not only the chemosensitivity of the neurons but also their morphological and physiological correlates, intracellular recording combined with dye injection of the neuron is necessary. Such an approach may be feasible in the gustatory neurons of the urodeles. In these animals, the chemosensitivity of their primary afferents was studied with extracellular recordings from the glossopharyngeal nerve (nIX) (Samanen and Bernard, '81; McPheeters and Roper, '85; Nagai, '891, which, together with the facial nerve (nVII) and the vagus nerve (nX), constitutes three pairs of cranial gustatory nerves. As to the morphological correlates, the central projection of these gustatory nerves was studied in the brainstem (Opdam and Nieuwenhuys, '76; Roth and Wake, '85). In the sensory ganglion or ganglia, however, the morphology of the gustatory neurons is not known. In light of this situation, we designed the present experiment to obtain fundamental information on the morphology of the glossopharyngeal nerve in the urodeles by applying a sensitive tracer to this nerve. We used the Mexican salamander, axolotl, Ambystoma mexicanurn, because the axolotls are available from inbred strains (Malacinski and Brothers, '74) and are easy to breed in a laboratory, and we can choose animals at desirable develop-

mental stages. We used labelling with cobaltic-lysine complex (Co-lys), since this method has proved effective in amphibians (Oka et al., '87). We studied the morphology of the cell bodies and their distribution in the ganglion of the IX nerve, and the projection of afferent fibers and the distribution of efferent cell bodies in the brainstem. Our results provide a morphological basis for future electrophysiological experiments on this nerve.

MATERIALSANDMETHODS Animals andsurgery Larvae of the axolotl (Ambystoma mexicanurn) were kindly provided by the Indiana University Axolotl Colony (Department of Biology, Bloomington, IN), and were maintained in our laboratory until approximately 1 year of age. Several phenotypes such as albino, white, and wild are established in the inbred strains of the axolotl (Malacinski and Brothers, '74). We used a white strain in the present experiments. Animals were anesthetized in 0.2% MS222 (tricaine methanesulfonate; Sankyo) for 20-30 minutes, laid in an operating chamber, and immersed in water except for the head region. Figure 1 shows the brain and some of the peripheral nerves in the axolotl. In amphibians, the glossopharyngeal (nIX), vagus (nX), and spinal accessory nerves form a complex of roots and the combined glossopharyngeal-vagus ganglion, and the configuration of this structure varies according to species (Roth and Wake, '85; Roth et al., '88). The axolotl (Opdam and Nieuwenhuys, '761, as well as the tiger salamander, Ambystoma tigrinurn (Herrick, '48),and the mudpuppy, Necturus (Wischnitzer, '791, lack the accessory nerve, but the IX and X nerves similarly form the combined ganglion. Thus, this ganglion is referred to as the IX-X ganglion in the present experiment. The ganglion gives off two roots to the brainstem: the rostral root and the caudal root of the IX-X ganglion. The caudal root consists of three rootlets (rt-1,2,3).The rostral root has two rootlets, which enter the brainstem dorsally and ventrally, respectively (see inset in Fig. 1).The ventral rootlet (rl-v) of the rostral root enters the brainstem slightly caudal to the dorsal rootlet (rl-d). However, these two rootlets run close together and are seen as a single bundle from a dorsal view of the brain. Therefore, the rostral root of the IX-X ganglion in Figures 5 and 12 is depicted as a single root. The IX-X ganglion gives off two peripheral branches of the IX nerve; a pretrematic branch to the facial nerve (R. communicans c.

Abbreviations

ax-IXen fsol nVIII nIX

nx

rl-d rl-v rl-1,2,3 R. comm . c. nVII tspv IXen IX-x

axon of efferent neuron of glossopharyngeal nerve solitary fasciculus octaval nerve glossopharyngeal nerve vagus nerve dorsal rootlet of rostral root of the IX-X ganglion ventral rootlet of rostral root of the IX-X ganglion three rootlets of caudal root of the IX-X ganglion Ramus communicans to the facial nerve spinal tract of trigeminal nerve cell body of efferent neuron of glossopharyngeal nerve glossopharyngeal and vagus complex

Fig. 1. Lateral view of t h e brain of Ambystoma mexicanurn. The white bars indicate t h e site of application of t h e tracers t o the Inset shows glossopharyngeal nerve (nIX) a nd t h e vagus nerve the details of rootlets of t h e IX-Xganglion.

(a).

GLOSSOPHARYNGEAL NERVE IN AXOLOTL

475

N. facialis) and a posttrematic branch (ramus posttrematicus) (Francis, '34; see also Samanen and Bernard, '81, for mudpuppy). The lingual branch of the ramus posttrematicus of the glossopharyngeal nerve was exposed just anterior to the first gill where it traverses the first epibranchial cartilage (Samanen and Bernard, '81; McPheeters and Roper, '85). Close to the first epibranchial cartilage, the ramus posttrematicus of the IX nerve of the axolotl gives off a branch to the gill area. Proximal to this branch and just distal to the branching point to the R. communicans c. N. facialis, the ramus posttrematicus was cut to apply the tracer (see Fig. 1).Note that the R. communicans c. N. facialis remained intact, so that the tracer was not applied to this branch. A Co-lys solution (Co-lys, Wako Chemical) containing 2-5% dimethyl sulfoxide (DMSO) was applied, by using a polyethylene tube, to the peripheral cut end of the IX nerve. The tube was sealed and secured to the nerve with a mixture of Vaseline and liquid parafin. After the Co-lys solution was applied to the tube, the animal remained anesthetized for additional 3-5 hours as a consequence of applying the MS222 solution onto the gills. This procedure facilitated incorporation of the tracer from the cut end. In the IX-X ganglion, the peripheral and central axons and the cell body of the IX afferents were labelled with Co-lys incorporated from the peripheral cut end of the IX nerve in a total of 14 axolotls. In the brainstem of the same preparations, the central axons of the afferents and the cell bodies of the efferents with dendrites were also labelled. In an additional three animals, the cell bodies of the IX nerve and the X nerve afferents were simultaneously labelled with HRP and Co-lys, respectively. In this experiment, Co-lys was applied to the peripheral cut ends of the four major branches, which the IX-X ganglion gives off caudally to the IX nerve (Fig. I). A 20% HRP solution containing 2-5% DMSO was applied to the IX nerve in a similar manner to the Co-lys labelling.

After 12 hours to 4 days survival, the animals were reanesthetized with MS222 and perfused transcardially, and the ganglia and the brain were processed for Co-lys. When both the IX nerve and the X nerve were cut for the HRP and Co-lys double labelling, the animals did not survive more than 2 days. Hence, for the double labelling the animals were killed after 24 hours survival. The histological procedures for Co-lys was basically after Antal ('851, and those for the HRP and Go-lys double labelling were after Toth and Szabo ('86) (see also Matsushima et al., '87, '88). Serial sections (50 km) of the tissue were cut. Horizontal sections were cut from the brainstem and ganglion in toto. For cross sections, the brainstem and the ganglion were severed in order to cut perpendicularly to the rostrocaudal axis of the brainstem and to the proximodistal axis of the ganglion, respectively. The sections were mounted on gelatinized slides, deparaffinized, and treated for CoS precipitate with Gallyas's developer for 20 minutes at 20°C for intensification. In some preparations a reducing agent (chloroauric acid) was eliminated from the developer to further intensify the end-products of Co-lys, although this procedure sacrificed a clear background in the section (see Antal, '85). The horizontal and the cross sections of the ganglion were made in five animals each, and the contours of cell bodies were traced with a camera lucida a t x 175. On the tracings, the major and minor diameters of the cell bodies were measured by using a digitizer connected to a personal computer, and the distribution of cell bodies within the ganglion was studied. The overall configuration of the IX-X ganglion was reconstructed from serial cross sections. For the brainstem, cross sections were made in five animals, and horizontal sections were made in three animals. The distributions of the afferent nerve fibers and the cell bodies

Fig. 2. Photomicrographs of Co-lys-labelled nIX cell bodies in the IX-Xganglion. Counterstained with Kernechtrot. A Horizontal section showing the general distribution ofcell bodies. Distal is to the right, and rostra1 is to the bottom. Note that the cell bodies are distributed

rostrally, but diffusely. Dark staining at bottom right shows labelled fibers of nIX (arrowheads). Scale bar = 100 pm. B: Higher magnification of the cell body marked by arrow in A. A single process leaves the cell body. Scale bar = 20 bm.

Tissue processing

476

T. NAGAI AND T. MATSUSHIMA

-aJ

u)

I

4

1

V

50-

-

I

E3

z

50pm Fig. 3. Camera lucida drawings of labelled IX nerve ganglion cells with a process.

of the efferent neurons were studied with serial cross sections. The rostral and caudal ends of their distributions were measured as a distance from the obex. The measured distances were averaged in five animals and are illustrated in a schematic diagram of the brainstem (Fig. 12). No correction for section shrinkage was applied to the brainstem or to the ganglion.

RESULTS Co-lyslabellingin the ganglion Go-lys labelled the cell bodies of the IX nerve as well as the fibers with dark-brown end-products (for their color, see Fig. 6A). Figure 2A shows the labelled cells and fibers in the IX-X ganglion sectioned horizontally. Figure 2B is an enlarged view of the cell body indicated by an arrow in A, showing a single process originating from the cell body. In the other cells in A, however, the processes are not clear. Although only a few cells with clearly labelled processes were found, they all possessed a single process. The morphology of these cells is illustrated in the camera lucida drawing of the representative cells (Fig. 3). The number of labelled cells in a ganglion varied between preparations, and ranged from 77 to 446 (ten ganglia). The major and minor diameters of the cell bodies were measured for all the ganglia sectioned horizontally (in 1,627 cells). The cell bodies were 22.7 pm (range: 6-58 pm) in the major diameter and 15.8 pm (range: 6-42 pm) in the minor diameter. Figure 4 shows the distribution profiles of the major diameter for five ganglia measured on horizontal sections. The distribution was unimodal. The cell size and the distribution profile were also examined in cross-sectioned ganglia, and the results were basically similar to those in the horizontal sections.

Distributionof the M nerve cell bodies in the ganglion Horizontal sections of the IX-X ganglion showed that the labelled cells were diffusely distributed in the rostral part of the ganglion (Fig. 2A). To examine the intraganglionic distribution of the cell bodies in greater detail, serial cross sections of 50 pm thick were made, and the location of the

n"

0-

I

I

Cell body diameter

I

50 pm

Fig. 4. Distributions of the size of the cell bodies. Major diameters are measured in five ganglia (different ganglia are marked with different symbols) sectioned horizontally. Arrow shows the mean diameter of total 1,627 cells.

cells was indicated by dots on the reconstructed ganglion (Fig. 5). The ganglion illustrated in Figure 5 had the most labelled cells among five ganglia examined. The general distribution patterns of the labelled cells in other ganglia were similar to the pattern in Figure 5B. The number of labelled cells in each section was counted, and the distribution of the cells of the IX nerve along the proximodistal axis of the ganglion was shown as a histogram (Fig. 5B, below). The histogram in Figure 5B shows no obvious mode on the longitudinal axis. Although it may appear to be of two modes in this particular ganglion, the distribution of the histogram was flatter with no indication of the mode in four other ganglia. Therefore, in general, the labelled cells were distributed uniformly along the proximodistal axis of the IX-X ganglion, with a slight concentration in its rostral part. However, some cells were also found in the caudal part, where the X nerve fibers were expected to pass through. To examine possible overlap in the distribution of the IX and X nerve components, we applied a double labelling method to the IX-X ganglion. Figure 6 shows a horizontal section of the IX-X ganglion simultaneously labelled with Go-lys and HRP. The ganglion cells labelled with either tracers were clearly distinguished by the different colors of the end-products of the tracers: the blue-black HRP end-products for the IX nerve and the dark-brown Co-lys end-products for the X nerve (Fig. 6A). The cell bodies of the X nerve labelled with Go-lys were widely distributed in the caudal as well as in the rostral part of the IX-X ganglion. The cell bodies and fibers of the IX

Fig. 5. Location of labelled cells in the IX-X ganglion. A: Examples of camera lucida drawings of cross sections. a, b, and c denote corresponding positions in B. Dorsal is to the right. Rostral is to the bottom. Note that the Co-lys-labelled cells are distributed diffusely. Scale bar = 500 pm. B: A semi-three-dimensional representation of a left IX-X ganglion. Dots show the location of the labelled cells. The dots in each cross section are projected onto the two-dimensional plane with a longitudinal shift every 50 pm, corresponding to the actual distance between cross sections. The contour of each section is shown every other section, namely a t 100 km, for clarity. The labelled cells are distributed mainly in the rostral part of the ganglion. Note that some labelled cells are also seen in the caudal part. The distribution of labelled cells along the proximodistal axis of the ganglion is shown in the histogram below, which is made at 50 pm intervals. There is no obvious mode in the histogram. Abscissa: The distance from the most proximal end of the ganglion. Ordinate: The number of labelled cells.

A

c

,

8

I I

0

I

/

.

...

1m 0

/

\

. .

,

\ \

IJm

Distance Figure 5

1000

Figure 6

GLOSSOPHARYNGEAL NERVE IN AXOLOTL nerve were mainly seen in the rostral part of the ganglion (Fig. 6B). In addition, a few IX nerve cell bodies were also found in the caudal part (Fig. 6A, arrow). In the double labelling experiments, however, the HRP usually labelled many fewer cells in the IX nerve than the Co-lys. This difference may have been caused by a weaker retrograde transport or sensitivity of HRP than Co-lys, because HRP showed no IX nerve afferents in the brainstem, but Co-lys did in the same preparation. Thus, the results of the double-labelling experiments suggest that the cell bodies of the IX nerve and X nerve are distributed mainly in the rostral and the caudal parts of the IX-X ganglion, respectively, but that there is some overlap in their distribution. We also examined label in the roots of the IX-X ganglion from the brainstem, and found that the fibers of the IX and X nerves were intermingled in the roots. The rostral root of the IX-X ganglion was divided into the dorsal rootlet and the ventral rootlet (Fig. 1). Examination of serial cross sections through these rootlets showed that the Co-lys applied to the IX nerve heavily labelled the ventral rootlet (rl-v) (Fig. 9A). A horizontal section through the level of the dorsal rootlet (rl-d) and the rootlets of the caudal root (r1-1,2,3)showed labelling in these rootlets, although it was weaker in the former than in the latter (Fig. 7). Co-lys applied to the X nerve heavily labelled the fibers in three rootlets of the caudal root and also in the dorsal rootlet of the rostral root (data not shown). Hence, the IX nerve fibers enter the brainstem mainly through the rostral root ventrally (rl-v), although a few fibers do so dorsally (rl-d). In addition, an entrance to the brainstem is provided by the caudal root which provides an entrance also for the X nerve. The detail in assigning these rootlets (rl-d,rl-v, and r1-1,2,3) to the IX and the X nerve components awaits further studies by a double labelling method.

Central projectionof afferent nerve When Co-lys was applied to the IX nerve, two sets of the root of the IX-X ganglion were labelled, i.e., the rostral root and the caudal root. The heavily labelled afferent fibers entering the brain through a ventral rootlet of the rostral root (rl-v) allowed us to follow their pathway in the brain (Figs. 9A, 10). The labelling in a rootlet of the caudal root (rl-1) was not heavy enough to follow its pathway in the brain (Fig. 7). The rostral afferent fibers entered the brainstem dorsal to the efferent fibers and formed the solitary fasciculus (Figs. 9A, 10). Rostrally, the afferent fibers almost reached the level of the cerebellum, and caudally, they could be followed to a level just rostral to the obex (Figs. 7, 8, 10, 12). The photomicrograph in Figure 9B corresponds to Figure 8C, and was taken at a level just caudal to the entrance of the rostral root, where the

Fig. 6. Horizontal section through the IX-X ganglion studied by double labelling of the glossopharyngeal nerve (1x1 with HRP and the vagus nerve (XI with Co-lys. A: Color photomicrograph showing the area divided by parallel lines in B. The cell bodies (arrowheads, arrow) and fibers labelled with HRP are clearly distinguished by its blue-black end-products. Note that the IX nerve cell in the caudal part of the ganglion (arrow) borders the X nerve cells, suggesting the IX and the Xnerve cells are somewhat interspersed in the IX-X ganglion. Scale = 100 pm. B: Camera lucida drawing of the horizontal section, which was taken 650 p m from the dorsum of the 800 km-thick ganglion. The cell bodies of the IX nerve are shown by the open tracings, while those of the IX nerve are shown by the filled tracings. The nerve fibers are traced only for the IX nerve. Scale = 500 pm.

479

Fig. 7. Horizontal section through the brainstem and the rootlets of the IX-X ganglion showing labelled nerve fibers in the rootlet of the caudal root (rl-1).Co-lys was applied to the IX nerve. Rostrally directed afferent fibers (fsol)and axons of the efferent neuron (ax-IXen)are also shown. Rostra1 is to the right. Lateral is to the bottom. Scale = 200 km.

labelling was most intense and the fibers were most tightly packed. Caudal to this level, the fasciculus became wider, but was subdivided into two or three small bundles (Fig. 8D-F, see also Fig. 11A). The labelled afferent fibers were also seen in the spinal tract of the trigeminal nerve (Figs. 9C, 10). These fibers ran caudally beyond the level of the obex (Fig. 12).A few rostrally directed fibers were also seen in the spinal tract of the trigeminal nerve; however, the labelling disappeared before it reached the level of the octaval nerve (nVIII) (Figs. 8B, 12). All label appeared in the brainstem ipsilateral to the Co-lys application.

Morphology of efferentneurons Efferent neurons in the IX nerve with developed dendritic arbors were seen in the brainstem ipsilateral to the Co-lys application (Fig. 11A). The cell bodies of the neurons were located near the border between the gray and the white matter (Fig. llA), and formed a longitudinal cell column (Fig. 11B). The column lay between the bulbar entrances of the rostral and the caudal roots of the IX-X ganglion (Fig. 12). The maximum number of the neurons in one preparation was 38 (mean = 27, in three horizontally sectioned preparations). The cell bodies of these neurons were roughly round with a major diameter of about 25 km. Their dendrites extended vetrolaterally to reach the ventral surface of the brainstem. Their axons ran ventral to the afferent fibers and left the brainstem through a ventral rootlet of the rostral root (rl-v) (Fig. 9A, see also Fig. 7). Figure 12 (lower half) schematically illustrates the distributions of the afferent fibers and the efferent neurons reconstructed from the serial frontal sections shown in Figure 8. The pattern of distribution averaged from five animals is also shown (Fig. 12, upper half ).

DISCUSSION Organizationof the M-X ganglion The present study examined for the first time the IX-X ganglion in a urodele by using retrograde labelling with

Fig. 8. Camera lucida drawings of a representative series of frontal sections through the brainstem and the upper spinal cord showing the labelled afferent and efferent components of the IX nerve. The rostrocaudal levels of the sections are indicated on a dorsal view of the axolotl brainstem in the lower right.

GLOSSOPHARYNGEAL NERVE IN AXOLOTL

Fig. 9. Photomicrographs of frontal sections of the brainstem. A A section at a level corresponding to Figure 8B, showing the entrance of the IX nerve through the ventral rootlet of the rostral root (rl-v). The afferent fibers enter the brainstem dorsally to the efferent fibers (arrow) to form the solitary fasciculus. Note no labelling in the dorsal

48 1

rootlet (rl-d). B: Enlarged view of the solitary fasciculus sectioned at a level corresponding to Figure 8C. C: Co-lys-labelled fibers in the descending trigeminal tract at a level corresponding to Figure 8E. Scale = 100 pm for A, 50 pm for B and C.

Fig. 10. Photomicrograph of a horizontal section of the brainstem. Dark staining in the ependyma and the lateral edge of the brainstem is an artifact. Note the ascending and descending fibers in the solitary fasciculus and the descending fibers in the spinal tract of the trigeminal nerve. Scale = 200 pm.

Co-lys. The organization of the ganglion was somewhat different from that in other vertebrates. In mammals, the IX nerve forms two separate sensory ganglia before it enters the brainstem. The ganglion located distally is called the petrosal or inferior ganglion, and the other located proximally the superior ganglion or ganglion of Ehrenritter (Clark, '26; Cajal, '52; Carmel and Stein, '69). The two ganglia are also identified in the IX nerve of avians (Dubbeldam et al., '79). The IX nerve in frogs forms two ganglia, but with a different terminology; the distal one is called the glossopharyngeal (1x1ganglion, and the proximal one the

jugular ganglion, which forms a complex with the vagus (X) (Gaupp, 1896-1904). In the bullfrog, these two ganglia of the IX nerve were quantitatively examined with retrogradely transported HRP by Hanamori and Ishiko ('83). They showed that the labelled cell bodies formed two separate masses in the rostral part of the IX-X ganglion; 62% of the total labelled cells were distributed in the IX nerve ganglion (distal ganglion), and 38% in the jugular ganglion. In the axolotl, however, the labelled cells were not seen in separate masses, but distributed diffusely as a single mass (Fig. 5). Moreover, some cell bodies of the IX nerve

482

T. NAGAI AND T. MATSUSHIMA

Fig. 11. Efferent neurons of the IX nerve labelled with Co-lys. A Frontal section at a level corresponding to Figure 8D. Note the round-shaped cell bodies (arrows) with well-developed dendritic arbors.

Scale = 100 km. B: Horizontal section of the brainstem showing a population of the efferent neurons distributed in a longitudinal cell column. Caudal is to the right. Lateral is to the bottom. Scale = 200 pm.

Fig. 12. Schematic dorsal view of the lower brainstem of the axolotl, showing the distribution patterns of the afferent nerve fibers and the efferent neurons of the IX nerve. The solitary fasciculus and the spinal tract of the trigeminal nerve are indicated by hatched and stippled areas, respectively. The location of the efferent neurons is indicated by dots. The distribution patterns in the left half of the brainstem are

reconstructed from the serial frontal sections of a representative single animal. In the right half, the averaged distribution patterns for the solitary fasciculus (hatched), the spinal tract of trigeminal nerve (stippled), and the efferent neurons (large dots) are shown. The bars indicate variation. The obex is used as a reference point. See Methods for details.

were also spread out in the caudal part of the IX-X ganglion, where the cell bodies and fibers of the X nerve were distributed. Our double labelling experiments with HRP and Co-lys confirmed that the distribution of the cell bodies of the IX and the X nerves overlapped (Fig. 6). Such an

overlap is not seen in mammals and frogs. In mammals, the two ganglia of the IX nerve are separated from the ganglia of the X nerve (nodose and jugular). In the combined ganglion of the IX and the X nerves of frogs, the distribution of the IX nerve cell bodies and fibers is demarcated

GLOSSOPHARYNGEAL NERVE IN AXOLOTL

483

from the X nerve component (Fig. 2C of Hanamori and Ishiko, '83). Consequently, the IX-X ganglion in the axolotl is relatively undifferentiated even compared with frogs, Because of the higher sensitivity of the cobaltic-lysine method (Oka et al., '871, the details of the cell body as well as the process in the axolotl IX nerve (Fig. 2B, 3) were more clearly seen than the HRP-labelled cells in the frog jugular ganglion (see Hanamori and Ishiko, '83). A single process originated from the successfully labelled cells. A bifurcation of the process, probably located far from the cell body, was not observed. However, we could assume that the cell with a single process is of the sensory pseudo-unipolar type, if it is a gustatory neuron, because it should give off a peripheral axon to the tongue and a central axon to the brainstem. Physiological experiments may elucidate this point. There may be other types of cells in the IX nerve in the axolotl, because sensory ganglion cells are not always pseudounipolar. In the auditory cranial nerve (the VIII nerve), bipolar and unipolar cells coexist in the spiral ganglion (Spoendlin, '73). In the nodose ganglion for the X nerve, a pseudo-bipolar cell is present as well as pseudo-unipolar ones (Helke and Hill, '88). The cell body size of the IX nerve in the axolotl (23 pm on average) may be equal to or smaller than that of the primary gustatory neurons in other species (for the IX nerve in dog, Clark, '26; monkey, Carmel and Stein, '69; cat, Berger, '80; frog, Hanamori and Ishiko, '83; for the VII nerve in rat, Miller et al., '78; chicken, Gentle and Hunter, '87). The IX nerve innervating the frog tongue contains at least two groups of functionally different afferents: mechanosensitive fibers with faster conduction velocities (C.V.) and chemosensitive fibers with slower C.V. (Hanamori and Ishiko, '81). In extracellular recordings from the axolotl IX nerve, the amplitude of spikes was larger following mechanostimulation than following chemo-stimulation, suggesting that fibers with different diameters exist in the IX nerve (Nagai, '89). Examination of cell body sizes of the IX nerve may reveal two groups of cells, since there is a positive correlation between fiber diameter and cell size (Lieberman, '76). However, the distribution of cell body diameters was unimodal (Fig. 4B), and did not suggest multiple cell groups. The IX nerve cells of different function and morphology may be shown by further studies with intracellular recordings combined with dye injection. In fact, the IX nerve ganglion cells in the axolotl permit stable impalements with intracellular microelectrodes (Nagai, unpublished data).

caudal edge of the cerebellum (Hanamori and Ishiko, '83; Stuesse et al., '84). At the level just caudal to the entrance of the rostral root of the IX-X ganglion, the intensely labelled fibers with Co-lys filled almost the entire solitary fasciculus. However, the cells which should be designated as the nucleus of the solitary fasciculus were not clearly seen in our counterstained preparations. In addition, caudal to that level, the labelled fibers did not form a single bundle (Fig. 8). Thus, the localization of the IX nerve afferents within the solitary fasciculus, as was suggested in frogs (Hanamori and Ishiko, '83; Stuesse et al., '84), was not clear in the axolotl. Afferent fibers in the IX nerve in the axolotl joined the spinal tract of the trigeminal nerve (nV) also, as was observed in the other species of vertebrates (monkey, Rhoton et al., '66; Beckstead and Norgren, '79; cat, Kerr, '62; rat, Torvik, '56, Contreras et al., '82; frog, Hanamori and Ishiko, '83, Stuesse et al., '84). The fibers descended along the lateral edge of the brainstem (Fig. 12), in a basically similar manner to those labelled with HRP in the frog (Fig. 7 of Hanamori and Ishiko, '83). However, the ascending fibers in the tract were much shorter in the axolotl.

Projectionof afFerent fibers

We gratefully acknowledge the continuous supply of axolotls from The Indiana University Axolotl Colony. We thank Dr. Y. Oka, Zoological Institute, Faculty of Science, University of Tokyo, for his critical reading of the manuscript, and Dr. P.A. Fuchs, Department of Physiology, University of Colorado School of Medicine, for improving the English in the revised manuscript. This study was supported by Grants-in-Aidfor Scientific Research (C) (No. 63540587) to T.N., and for Encouragement of Young Scientists (No. 63740408) to T.M., from the Ministry of Education, Science, and Culture of Japan.

Herrick ('44) pointed out that in Ambystoma tigrinum all visceral afferent components of the cranial nerve converge into the solitary fasciculus. As for the cranial nerve in the axolotl (Ambystoma mexicanum), Opdam and Nieuwenhuys ('76) made only a general description on the cytoarchitecture of the brainstem by Nissl and Bodian staining. They showed in their topological analysis that the visceral afferents (the VII, IX, and X nerves) converge into the solitary fasciculus. In the present study, selective labelling 3f the IX nerve with Co-lys allowed us to delineate its projection area in the brainstem (Fig. 12). The afferents of the IX nerve bifurcate in the brainstem, and the ascending component runs more rostrally than the visceral afferents in Opdam and Nieuwenhuys's analysis (Opdam and Nieuwenhuys, '76). Similar rostral extension is also observed in frogs, in which the ascending component of the IX nerve reaches the

Efferent neurons in the brainstem By selectively labelling with Co-lys, the efferent nucleus of the IX nerve was delineated in the brainstem more specifically (Fig. 12) than by Opdam and Nieuwenhuys's analysis (Opdam and Nieuwenhuys, '76), in which the efferent nuclei of the IX and X nerves are shown as a single cell mass. The efferent nucleus in the axolotl was more longitudinally elongated than that in the frogs (Hanamori and Ishiko, '83; Stuesse et al., '84) and the toad (Oka et al., '871, as was observed in other salamanders (Roth and Wake, '85; Roth et al., '88). The diameter and the number of the IX nerve efferent neurons in the axolotl were similar to those in salamanders (21 pm; Opdam and Nieuwenhuys, '76; 25-40 neurons; Roth et al., '88).The efferent neurons in the axolotl seemed to be of a single population, judging from their size and the morphology of cell body and dendrites. Smaller cells with different morphology, suggestive of preganglionic parasympathetic neurons, as have been reported in the toad (Oka et al., '871, were not observed in our preparation.

ACKNOWLEDGMENTS

Akaike, N., A. Noma, and M. Sat0 (1976) Electrical responses of frog taste cells to chemical stimuli. J. Physiol. (Lond.) 254:87-107. Antal, M. (1985) The cobalt as a neuronal tracer. Acta Med. Kinki Univ. 1O:l-10.

484 Attwell, D., P. Mobbs, M. Tessier-Lavigne, andM. Wilson (1987) Neurotransmitter-induced currents in retinal bipolar cells of the axolotl, Ambystoma mexicanum. J. Physiol. (Lond.) 387:125-161. Avenet, P., F. Hofmann, and B. Lindemann (1988) Transduction in taste receptor cells requires CAMP-dependent protein kinase. Nature 331:351354. Beckstead, R.M., and R. Norgren (1979) An autoradiographic examination of the central distribution of the trigeminal, facial, glossopharyngeal, and vagal nerves in the monkey. J. Comp. Neurol. 184:455-472. Berger, A.J. (1980) The distribution of the cat’s carotid sinus nerve afferent and efferent cell bodies using the horseradish peroxidase technique. Brain Res. 190:309-320. Cajal, S.R. (1952) Histologie du Systeme Nerveux de 1’Homme e t des Vertebres, Vol. I. Madrid: Consejo Superior de Investigaciones Cientificas. Carmel, P.W., and B.M. Stein (1969) Cell changes in sensory ganglia following proximal and distal nerve section in the monkey. J. Comp. Neurol. 135:145-166. Clark, S.L. (1926) Nissl granules of primary afferent neurones. J. Comp. Neurol. 41r423-452. Contreras, R.J., R.M. Beckstead, and R. Norgren (1982) The central projections of the trigeminal, facial, glossopharyngeal and vagus nerves: An autoradiographic study in the rat. J. Auton. Nerv. Syst. 6r303-322. Delay, R.J., and S.D. Roper (1988) Ultrastructure of taste cells and synapses in the mudpuppy Necturus maculosus. J. Comp. Neurol. 277268-280. Dennis, M.J., and J.W. Yip (1978) Formation and elimination of foreign synapses on adult salamander muscle. J. Physiol. (Lond.) 274299-310. Diamond, J., E. Cooper, C. Turner, and L. Macintyre (1976) Trophic regulation of nerve sprouting. Science 193:371-377. Dubbeldam, J.L., E.R. Brus, S.B.J. Menken, and S. Zeilstra (1979) The central projections of the glossopharyngeal and vagus ganglia in the mallard,Anas platyrhynchos L.J. Comp. Neurol. 183:149-168. Farbman, A.I., and J.D. Yonkers (1971) Fine structure ofthe taste bud in the mud puppy, Necturus maculosus. Am. J. Anat. 131:353-370. Francis, E. (1934) The Anatomy of the Salamander. Oxford: Clarendon Press. Gaupp, E. (1896-1904) Anatomie des Frosches. Braunschweig: Friedrich Viewegund Sohn. Gentle, M.J., and L.N. Hunter (1987) The geniculate ganglion of the chicken (Gallusgallus war domesticus). Med. Sci. Res. 15:637-638. Getchell, T.V., and G.M. Shepherd (1978) Responses of olfactory receptor cells to step pulses of odour at different concentrations in the salamander. J. Physiol. (Lond.) 282:521-540. Hanamori, T., and N. Ishiko (1981) Conduction velocity of the IXth nerve fibers innervating taste organs in the rostral and caudal tongue region in bullfrog. Chem. Senses 6:175-187. Hanamori, T., and N. Ishiko (1983) Intraganglionic distribution of the primary afferent neurons in the frog glossopharyngeal nerve and its transganglionic projection to the rhombencephalon studied by HRP method. Brain Res. 260:191-199. Helke, C.J., and K.M. Hill (1988) Immunohistochemical study of neuropeptides in vagal and glossopharyngeal afferent neurons in the rat. Neuroscience 261539-551. Herrick, C.J. (1944) The fasciculus solitarius and its connections in amphihians and fishes. J. Comp. Neurol. 81:307-332. Herrick, C.J. (1948) The Brain of the Tiger Salamander. Chicago: Univ. of Chicago Press. Kauer, J.S. (1988)Real-time imaging of evoked activity in local circuits of the salamander olfactory bulb. Nature 331:166-168. Kerr, F.W.L. (1962) Facial, vagal and glossopharyngeal nerves in the cat. Arch. Neurol. 6.264-281. Kinnamon, S.C., and S.D. Roper (1987) Passive and active membrane properties of mudpuppy taste receptor cells. J. Physiol. (Land.) 383:601614. Kinnamon, S.C., and S.D. Roper (1988) Membrane properties of isolated mudpuppy taste cells. J. Gen. Physiol. 91r351-371. Kuffler, S.W., J.G. Nicholls, and R.K. Orkand (1966) Physiological properties of glial cells in the central nervous system of amphibia. J. Neurophysiol. 29:768-787. Lieberman, A.R. (1976) Sensory ganglia. In D.N. Landon (ed): The Peripheral Nerve. London: Chapman and Hall, pp. 188-278. Malacinski, G.M., and A.J. Brothers (1974) Mutant genes in the Mexican axolotl. Science 184:1142-1147.

T. NAGAI AND T.MATSUSHIMA Matsushima, T., M. Satou, and K. Ueda (1987) Direct contacts between glossopharyngeal afferent terminals and hypoglossal motoneurons revealed by double labeling with cobaltic-lysine and horseradish peroxidase in the Japanese toad. Neurosci. Lett. 80:241-245. Matsushima, T., M. Satou, and K. Ueda (1988) Neuronal pathways for the lingual reflex in the Japanese toad. J. Comp. Physiol. [A.] 164:173-193. McPheeters, M., and S.D. Roper (1985) Amiloride does not block taste transduction in the mudpuppy, Necturus maculosus. Chem. Senses 10:341-352. Miller, I.J., Jr., M.M. Gomez, and E.H. Lubarsky (1978) Distribution of the facial nerve to taste receptors in the rat. Chem. Senses 3r397-411. Model, P.G. (1978) Aspects of Mauthner cell differentiation in the axolotl, Ambystoma mexicanum. Am. Zool. 18253-265. Nagai, T. (1989) Chemoresponses of the glossopharyngeal nerve of the axolotl (Ambystoma mexicanum). Chem. Senses 14:315 (Abstract). Oka, Y., H. Takeuchi, M. Satou, and K. Ueda (1987) Cobaltic lysine study of the morphology and distribution of the cranial nerve efferent neurons (motoneurons and preganglionic parasympathetic neurons) and rostral spinal motoneurons in the Japanese toad. J. Comp. Neurol. 259:400423. Opdam, P., and R. Nieuwenhuys (1976) Topological analysis of the brain stem of the axolotl Ambystoma mexicanum. J. Comp. Neurol. 165285306. Ozeki, M. (1971) Conductance change associated with receptor potentials of gustatory cells in rat. J. Gen. Physiol. 58:688499. Rhoton, A.L., Jr., J.L. O’Leary, and J . P . Ferguson (1966) The trigeminal, facial, vagal, and glossopharyngeal nerves in the monkey. Arch. Neurol. 14:530-540. Roper, S. (1983) Regenerative impulses in taste cells. Science 220:13111312. Roper, S.D. (1989) The cell biology of vertebrate taste receptors. Annu. Rev. Neurosci. 12:329-353. Roth, G., and D.B. Wake (1985) The structure of the brainstem and cervical spinal cord in lungless salamanders (family plethodontidae) and its relation to feeding. J. Comp. Neurol. 241:99-110. Roth, G., K. Nishikawa, U. Dicke, and D.B. Wake (1988) Topography and cytoarchitecture of the motor nuclei in t h e brainstem of salamanders. J. Comp. Neurol. 278:181-194. Samanen, D.W., and R.A. Bernard (1981) Response properties of the glossopharyngeal taste system of the mud puppy (Necturus maculosus). J. Comp. Physiol. 143~143-150. Sato, M (1976) Physiology of the gustatory system. In R. Llinas and W. Precht (eds): Frog Neurobiology. Berlin: Springer, pp. 576-587. Sato, T., and L.M. Beidler (1975) Membrane resistance change of the frog taste cells in response to water and NaC1. J. Gen. Physiol. 66r735-763. Spoendlin, H. (1973) The innervation of the cochlear receptor. In A.R. MBller (ed): Basic Mechanisms in Hearing. New York Academic Press, pp. 185-234. Stuesse, S.L., W.L.R. Cruce, and K.S. Powell (1984) Organization within the cranial M-X complex in ranid frogs: A horseradish peroxidase transport study. J. Comp. Neurol. 222:358-365. Sugimoto, K., and J.H. Teeter (1987) Voltage-dependent and chemicallymodulated ionic currents in isolated taste receptor cells of the tiger salamander. Proc. Soc. Neurosci. 13.1404 (Abstract). Torvik, A. (1956) Afferent connections to the sensory trigeminal nuclei, the nucleus of the solitary tract and adjacent structures: An experimental study in the rat. J. Comp. Neurol. 106:51-141. Toth, P., and T. Szabo (1986) Simultaneous labelling of two different central nervous system pathways with horseradish peroxidase and cobalt in Gnathonemus petersii and Rana esculenta. Neurosci. Lett. 64:350-354. Toyoshima, K., K. Miyamoto, and A. Shimamura (1987) Fine structure of taste buds in the tongue, palatal mucosa and gill arch of the axolotl, Ambystoma mexicanum. Okajimas Folia Anat. Jpn. 64t99-110. Werblin, F.S., and J.E. Dowling (1969) Organization of the retina of the mudpuppy Necturus maculosus. 11.Intracellular recording. J. Neurophysiol. 32t339-355. West, C.H.K., and R.A. Bernard (1978) Intracellular characteristics and responses of taste bud and lingual cells of the mudpuppy. J. Gen. Physiol. 72: 305-32 6. Wischnitzer, S. (1979) Anatomy of the mudpuppy, Necturus. In D. Kennedy and R.B. Park (eds): Atlas and Dissection Guide for Comparative Anatomy. San Francisco: W.H. Freeman, pp. 73-105. Yang, J., and S.D. Roper (1987) Dye-coupling in taste buds in the mudpuppy, Necturus maculosus. J. Neurosci. 7:3561-3565.