To clarify when the free neuromasts of fish become functional, the neuromast formation in the embryonic development of the cod fish Gadus macrocephalus Tilesius was investigated. The initial appearance of placodes and an increase in the number and distribution of free neuromasts were followed using light microscopy and scanning electron microscopy. Subsequently, the appearance of afferent and efferent nerve endings of the free neuromasts was investigated by transmission electron microscopy. Sections of embryos 72 hr before hatching revealed a pair of placodes on the head and two pairs on the trunk. The placodes increased in number, and some of them differentiated into free neuromasts during the embryonic stage. There were three pairs of free neuromasts on the head and four pairs on the trunk in newly hatched larvae. The afferent nerve endings were recognized first in the free neuromasts on the trunk at about 36 hr before hatching, and the efferent nerve endings were recognized on the trunk at about 24 hr before hatching. On the head, afferent nerve endings were seen at about 12 hr before hatching; the efferent nerve endings were not yet seen in newly hatched larvae.
Free neuromasts are known as a type of fish lateral line organ that are formed on the head and body during the early developmental stages. They first appear as placodes and differentiated into free neuromasts. Some of them are incorporated into canals at the last larval stage.
There are many reports concerning the development, morphological changes, distribution and sensory cell polarity of the free neuromasts during the larval stage (Blaxter, 1984a,b; Blaxter et al., 1983; Cahn and Shaw, 1962; Harvey et al., 1992; Iwai, 1963b, 1967, 1972; Kawamura and Ishida, 1985; Kawamura et al., 1983; Metcalfe et al., 1985; Mukai and Kobayashi, 1995; Mukai et al., 1992; Sato, 1952). Some species of fish embryos are known to have some free neuromasts and cupulae on the head and body (Iwai, 1963a, 1964, 1965; Mukai and Kobayashi, 1993). However, few reports have described the initial formation of placodes and their subsequent differentiation as the embryo grows (Iwai, 1963a; Sato, 1952). To investigate the functions of free neuromasts, some investigators have used electrophysiology methods (Katsuki and Yoshino, 1950; Katsuki et al., 1950, 1951; Kawamura and Yamashita, 1983; Suga, 1967). Histologically, it is generally believed that neuromasts showing two types of nerve endings, afferent and efferent, are functional (Iwai, 1972). Metcalfe et al. (1985) reported on the nerve endings of the neuromasts in five-day-old zebrafish larvae. However, there are neither electrophysiological nor histological studies on the functions of free neuromasts in fish embryos.
This paper describes the results of an investigation into the morphology and function of free neuromasts in the developing cod embryo of Gadus macrocephalus Tilesius using light microscopy, scanning electron microscopy (SEM) and transmission electron microscopy (TEM).
MATERIALS AND METHODS
Artificially fertilized Gadus macrocephalus Tilesius eggs were obtained from the Japan Cultivation Fisheries Association of Noto Station, Ishikawa Prefecture. They were transferred to the laboratory aquarium where the water was kept at 9°C.
Embryos were fixed every 6–12 hr from 7 days after fertilization without being anesthetized in a mixture of 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1M phosphate buffer solution (pH. 7.4) for 1.5 hr at room temperature. They were rinsed with the buffer solution for more than 12 hr at 4°C. The egg membrane was carefully removed by forceps in the buffer solution under a binocular and post-fixed in 1% osmic acid for 2 hr at 4°C. For SEM observation, specimens were dehydrated in graded ethanols, dried at the critical point with liquid CO2, sputtered with gold and examined with a HITACHI S-600 microscope. For the TEM examination, specimens were dehydrated in graded acetones and embedded in an epoxy resin (Epon 812). A JEOL 200 CX microscope was used for observation. Semithin sections (about 1 μm) of the epoxyembedded specimens were also observed under a light microscope after staining with toluidine blue.
Embryonic stages are designated by hours before hatching (−72 hr, −48 hr etc.). The placodes and neuromasts are numbered in order of appearance, on the head (I, II, etc.) and on the trunk and tail (i, ii, etc.).
Embryonic development and formation of free neuromasts
The incubation period was about 240 hr, and the isolation of a tail from the yolk sac was recognized around 92 hr after fertilization. Table 1 shows the development of the free neuromast, and Figs. 1 and 2 show the distribution of the free neuromasts formed during the embryonic stage. In the embryos 72 hr before hatching (−72 hr), the protuberance of the embryonic body from the yolk sac is slight and the end of the tail reached to an eye by curling of the body. Lightcolored melanophores were distributed all over the body. The nasal cavities were seen, the lenses of the eyes were already developed, and pectoral fins were seen as bud-like processes. Sections showed that a pair of narrow cracks (gill holes) are open to the exterior below the auditory vesicles. At this stage there were a pair of placodes on the head, placode I, about 50 μm wide and 23 μm high, located upper-posterior to the eyes (Fig. 3A). On the trunk, there were two pairs of placodes: placode i, about 50 μm wide and 22 μm high, posterior to the pectoral fin and placode ii, about 38 μm wide and 20 μm high, at the middle of the trunk. These placodes showed cupulaholes. The diameter of the cupula-hole of placode I was 3 μm, of placode i, 7.5 μm, and of placode ii, 2.5 μm.
Free neuromast formation in Gadus macrocephalus embryos
At about −48 hr, the number of placodes increased (Table 1). On the head, in addition to placode I, there were another four pairs of placodes: placode II, between the nasal cavities; placode III, upward of the eyes; placode IV and placode V, anterior and posterior to the abdominal side of the eyes (Fig. 2). There were one or two more placodes, placode iii and placode iv, on either side of the trunk (Fig. 1).
At about −40 hr, placode i was differentiated into a free neuromast (Table 1) whose apical surface was visible by SEM and was about 7.5 μm in width, with a few kinocilia and stereocilia.
At about −36 hr, the tail curved over the trunk and head and its tip reached the pectoral fin which protruded semicircularly (Fig. 4A). The arrowhead of Fig. 4C shows a gill hole cracking inside the body, and its external appearance by SEM is shown by the arrowhead in Fig. 4A.
Twenty-four hr before hatching, the melanophores were distributed on the head and aggregated on the dorsal side of the digestive duct and two places of the tail. At this stage, placodes I and II on the head and placode ii on the trunk differentiated into free neuromasts (Table 1). The apical surface of neuromast I was caved in and was about 8 μm in width with five-to-six kinocilia about 4.7 μm long (Fig. 3C, C′). The diameter of the apical surface of neuromast II was 7.5 μm, and there were few cilia; the apical surface of neuromast i was about 10 μm wide, and nine-to-ten kinocilia about 3.6 μm long were seen.
Twelve hr before hatching, placode VI with a cupula-hole was seen behind the placode V on the head (Fig. 2). On the trunk, placode iii became visible by SEM.
Just after hatching, the body length of the larva was about 3.8 mm. On the head, only the top of the snout was isolated from the yolk sac (Fig. 4B). There were still only one pair of gill holes, but gills were observed in the sections cut at the level of the holes (Fig. 4B,D). An anus was formed. The pectoral fins became disc-like shapes (Fig. 4B). Newly hatched larvae had three pairs of free neuromasts and three pairs of placodes on the head (Fig. 2) and four free neuromasts and two placodes on either side of the trunk. Figure 3D shows a free neuromast projected from the epidermis with a flat apical surface. The cupula-holes had disappeared by this stage. The apical surface of a free neuromast detected by SEM is shown in Fig. 3 D′. Free neuromast i exhibited the largest apical surface, 12.5 μm in width.
The mouth was recognized as a thin groove in larvae about one day old. In about larvae two days old, with a body length of 4.5 mm, the movement of the mouth, twitching, could be seen. After hatching, the pigmentation of the eyes developed, but heavy black pigmentation was seen two days after hatching. The cupulae could be seen clealy in one-day-old larvae for the first time; it was about 35 μm long, present on free neuromast i.
TEM observation and formation of the nerve endings
At about −36 hr, a nerve bundle about 2.7 μm in diameter was in contact with a free neuromast. In free neuromast i, an afferent nerve ending was formed between the nerve cell and the sensory cell. A dense body about 250 μm in diameter was seen in the base of the sensory cell (Fig. 5A). The apical surface of neuromast I was caved in (cupula-hole) and filled with indeterminate formed substances (Fig. 5B). One kinocilium and at least 29 streocilia composed a group of cilia, and the kinocilium was located in the periphery of this group. There were at least five groups of cilia in one neuromast, and they were located in opposite directions from each other. There was a fine fibrous network (Flock, 1965) between the stereocilia (Fig. 5C).
At about −24 hr, in neuromast I, at least five kinocilia were recognized and an afferent nerve ending was also seen. An efferent nerve ending was observed in free neuromast i; there was an accessory double membrane at the sensory cell side, and the nerve cell adjacent to this double membrane contained synaptic vesicles (Fig. 5D).
At about −12 hr, afferent and efferent nerve endings were recognized in neuromast ii. The periphery of the nucleus of the sensory cells often shows deep recesses. The nucleus of the supporting cells arranged at the basement of the neuromast, and there were well-developed endoplasmic reticulum around the nucleus. In neuromast i, one kinocilium and at least 32 stereocilia comprised a group of cilia.
We observed that in Gadus macrocephalus, the placodes increased in number, and some of them differentiated into free neuromasts before hatching. The number of free neuromasts of newly hatched larvae depends on the species (Iwai, 1972; Mukai and Kobayashi, 1995). Gadus Morhua L. possess two pairs of free neuromasts on the head and five on either side of the trunk (Blaxter, 1984b), but G. macrocephalus has three pairs on the head and four on either side of the trunk. According to Blaxter (1984b), the proliferation of neuromast organs in G. morhua is slow; 40 days after hatching, the larvae still only have nine pairs on the trunk. Blaxter suggested that this is due to the more shortened body form of this cod larvae.
Three pairs of placodes with cupula-holes ware recognized in the −72 hr G. macrocephalus embryos. As shown in Table 1, placode i, which is located posterior to the pectoral fins on the trunk (Fig. 1) was the placode that was formed first and also differentiated into a free neuromast first. The time required for the differentiation from placode to free neuromast varied. Placode i took about 30 hr to differentiate into a free neuromast. Placode I, on the head, took about 48 hr to differentiate into a free neuromast. Neuromasts IV and V were formed on the abdominal side of the eyes and became visible by SEM after the yolk was absorbed. Neuromasts IV was seen in 1-day-old larvae. On the trunk, the proliferation of free neuromasts seemed to occur in sequence from anterior to posterior in proportion to the body growth.
Placodes are made in the epidermis, and there is a cupulahole at the center of the placodes (Fig. 3 A,B). The cupulahole is a hollow made as a result of the radially arranged cells of the placode (Iwai, 1963a) and filled with indeterminate formed substances (Fig. 5B). When the apical surface was open to the exterior, it was caved in and became visible by SEM (Fig. 3C,C′). By hatching, free neuromasts projected from the epidermis and their apical surface became flat (Fig. 3D,D′) and the cupula-holes disappeared. According to Iwai (1967), cupula-holes are seen for a short period before the cupula appears. The apical surface of free neuromasts of the two-day-old larvae is rather hemispherically projected.
In killifish medaka, Oryzias latipes, Sato (1952) described that in some cases one placode is differentiated into two or three neuromasts during the embryonic growth. We did not observe such a phenomenon in G. macrocephalus embryos.
There is general agreement that the formation of two kinds of nerve endings of the sensory organs indicates that the organs are functional (Iwai, 1967, 1972). The neuromasts in oneday-old medaka larvae (Iwai, 1964), and in five-day-old zebrafish larvae (Metcalfe et al., 1985) are suggested to be functional. Iwai (1964) also reported that these two types of nerve endings appear after the cupula formation is accomplished. Therefore, in G. macrocephalus, neuromast i seems to be functional at least at −24 hr, and neuromast ii is functional at about −12 hr. However, the cupulae were not observed until the larva was 24 hr old. They might have been lost due to handling, because they are easily broken. On the head, in contrast, efferent nerve endings were not seen during the embryonic stage. They seem to be formed by the first feeding stage of the larva (Blaxter and Fuiman, 1989). In G. macrocephalus embryos, the free neuromasts found on the trunk are functional earlier than those on the head.
The developmental stage at which free neuromasts would first be functional seems to differ among species. It probably depends on habitat and habits, especially just after hatching. The ayu Plecoglossus altivelis and cyprinid fish Gnathopogon elongatus caerulescens have well developed free neuromasts with cupula when they hatch. The former needs such well-developed neuromasts because just after hatching, they swim actively and downward into the sea. In the latter, which usually feed on animal plankton, it is thought that well-developed neuromasts are essential to detect the fine water flow produced by their prey. In contrast, another cyprinid fish, Zacco platypus, is hatched with undeveloped free neuromasts or eyes. They stay in the spawing bed at the bottom of the river about four days after hatching until these organs develop further. (Kawamura and Ishida, 1985; Kawamura et al., 1983).
In G. macrocephalus, we observed that during the embryonic development, there were a pair of ectodermal invaginations below the auditory vesicles, as indicated by the arrowheads in Fig. 4A and B. Sections cut at this level showed that these invaginations eventually meet the gill cavity (Fig. 4C,D). We coined the term “gill hole” to refer to them, since there is no previous report of these holes. We recognized the same invagination in other kinds of fish embryos (unpublished data). We suspect that they are involved with respiration, and are one of the important aspects of morphogenesis taking place in the fish embryo.
We are grateful to Prof. M. Tanaka of the Kyoto University and Mr. K. Yoseda of the Japan Cultivation Fisheries Association of Noto Station for kindly supplying the fertilized eggs.