Sensory neuroanatomy of a skin-penetrating nematode parasite:Strongyloides stercoralis. I. Amphidial neurons

THE JOURNAL OF COMPARATIVE NEUROLOGY 357~281-295 (1995)

Sensory Neuroanatomy of a Skin-Penetrating Nematode Parasite: Strongyloides stercoralis. I. Amphidial Neurons F.T. ASHTON, V.M. BHOPALE, A.E. FINE, AND G.A. SCHAD Department of Pathobiology, School of Veterinary Medicine, and NIH-IVEM Image Processing Resource, University of Pennsylvania, Philadelphia, Pennsylvania 19104

ABSTRACT The Strongyloides stercoralzs infective larva resumes feeding and development on receipt o f signals, presumably chemical, from a host. Only two of the anterior sense organs of this larva are open to the external environment. These large, paired goblet-shaped sensilla, known as amphids, are presumably, therefore, the only chemoreceptors. Using three-dimensional reconstructions made from serial electron micrographs, amphidial structure was investigated. In each amphid, cilialike dendritic processes of 11neurons extend nearly to the amphidial pore; a twelfth terminates at the base of the amphidial channel, behind an array of lateral projections on the other processes. A specialized dendritic process leaves the amphidial channel and forms a complex of lamellae that interdigitate with lamellae of the amphidial sheath cell. This “lamellar cell” is similar to one of the “wing cells” or possibly the “finger cell” of Caenorhabditis elegans. Each of the 13 amphidial neurons was traced to its cell body. Ten neurons, including the lamellar cell, connect to cell bodies in the lateral ganglion, posterior to the nerve ring. The positions of these cell bodies were similar to those of the amphidial cell bodies in C. elegans. Therefore, they were named by using C. elegans nomenclature. Three other amphidial processes connect to cell bodies anterior to the nerve ring; these have no homologs in C. elegans. A map allowing identification of the amphidial cell bodies in the living worm was prepared. Consequently, laser ablation studies can be conducted to determine which neurons are involved in the infective process. 199: Wiley-Liss, Inc. r)

Indexing terms: Cuenorhubditis elegans, dendrite, development, three-dimensionalreconstructions,

ultrastructure

Despite its central role in the establishment of parasitism, the infectious process of soil-transmitted parasitic nematodes (e.g., Strongyloides spp. and hookworms) is poorly understood. The signals that reactivate development in infective larvae (L3) following entry into the host are unknown. Free-living, infective larvae (L3) of Strongyloides stercoralis resemble the resting (dauer) stage larvae of Caenorhabditis elegans (Politz and Philip, 1992; Hotez et al., 1993). Both the L3 and the dauer larvae are environmentally resistant, developmentally arrested stages. S. stercoralis third-stage larvae recognize hosts, initiate penetration, and resume feeding and development after entry into a host. Similarly, C. elegans dauer larvae also resume feeding and development when new resources become available. As a basis for studies planned to elucidate the mechanisms of host invasion and, particularly, resumption of development in S. stercoralis, we have initiated a study of the sensory ultrastructure of the infective third-stage larva. To our B 1995 WILEY-LISS, INC.

knowledge, there has been no study of the sensory ultrastructure of an economically important nematode parasite comparable to the detailed analyses of the neuroanatomy of C. elegans (Ward et al., 1975; Ware et al., 1975; White et d., 1986). Only 2 of the 18 anterior sensilla of the infective thirdstage larva of s.stercoralis, namely the amphids, are open to the external environment. The amphids are large, lateral, paired sensilla, each containing a bundle of sensory neurons exposed to the outside through a goblet-shaped supporting cell, the sheath cell. Because the amphids are open, it is likely that the sensory neurons that recognize either entry into a host or in vitro hostlike conditions Accepted December 11, 1994. Addross raprint requests to G.A. Schad, Department of Pathobiology, School of Veterinaly Medicine, University of Pennsylvania, Philadelphia, PA 19104.

F.T. ASHTON ET AL.

282 (Hawdon and Schad, 1990) must be in these sensilla. This agrees with recent work on resumption of development in the dauer larva (a resting stage) of Caenorhabditis elegans, which demonstrated that interruption and resumption of development are under the control of amphidial neurons (Bargmann et al., 1990; Bargmann and Horvitz, 1991a,b). We have traced the 13 amphidial neurons, the putative main chemoreceptors, from the “nose” of the worm to their distant cell bodies. We have recorded their positions and those of other, as yet unidentified, nearby cell bodies to create a map for use in laser ablation studies that are planned to determine the function of the individual neurons.

MATERIALS AND METHODS Fixation Because infective stage larvae of S. stercoralis, like most nematode larvae, are highly resistant to conventional fixation methods, we have developed a procedure based on the method of Jones and Ap Gwynn (1991) that combines microwave radiation with conventional fixatives. For fixation, larvae were transferred to a 1.5-ml Eppendorf-type centrifuge tube containing 4% glutaraldehyde, buffered in 0.1 M cacodylate buffer, pH 6.8. The centrifuge tube was suspended in a 30-ml beaker containing approximately 25 ml of water at 21°C that was then placed in a microwave oven. The oven (General Electric Model JEM21) was operated at 20%’power for 2 minutes, causing the water in the beaker (and thus the fixative in the centrifuge tube) to reach 48-50°C. The worms were rinsed three times in buffer, microwaved in Os04,buffered as described earlier. After three rinses in water, the worms were exposed to microwaves a third time in 1%uranyl acetate in water. In our experience with S. stercoralis, as well as with hookworms, and the stomach worm of ruminants Haemonchus contortus, it seems necessary to apply the microwave energy in very short bursts. Otherwise, there appears to be localized heating of certain structures (possibly those with a large amount of either water or lipid), resulting in swelling of these structures and a distortion of those surrounding them. After two additional rinses in water, the anterior onethird to one-fourth of the worms was cut off with a diamond scalpel (Microstar)to facilitate dehydration and embedding. The worms were dehydrated in acetone and embedded in 1:l Araldite 502:DDSA (Polysciences, Inc.). Eight to 10 cycles of vacuum were used to assure complete infiltration of the plastic monomer. The plastic was polymerized overnight in a 60°C oven. Serial sections 0.12-0.25-pm thick were cut on a Servall MT-2 ultramicrotome fitted with a diamond knife. The ribbons of sections were picked up on Formvar-covered slotted grids. The sections were stained in 1% uranyl acetate in methanol, followed by Reynold’s lead citrate. Two worms were sectioned well past the nerve ring (650700 serial sections), and only occasionally was a single section lost. Two more complete sets of sections through the amphid (60-65 sections), as well as a number of less complete series, were also cut.

Electron microscopy The sections were examined and photographed in a JEOL JEM-4000EX Intermediate voltage electron microscope, operated at 300 KV. The grids were mounted in the

microscope’s rotation holder. Serial sections were generally photographed at a magnification of 8,000X . The images on the negatives were projected on the screen of a Bausch and Lomb optical comparator, and the contours of the neurons were traced onto transparent acetate sheets. Registration was determined by aligning each section with the tracings of the neurons in the previous section. Part of the pharynx was usually visible in the projected image; this, as well as muscle bundles, was used as a guide in alignment of the sections. In some of Ihe sections, the neurons were running obliquely, therefore their outlines were not entirely clear. After the specimen was rotated to find the orientation of the image where the outlines of the neurons were most clear, these sections were rephotographed in stereo at a magnification of 2 0 , 0 0 0 ~or 3 0 , 0 0 0 ~ .In the stereo images the outlines of the neurons could be seen clearly.

Computer imaging To make the three-dimensional (3-D) reconstructions of the amphidial neurons, the contours of the dendritic processes in each section were digitized by using the program SDED (Young et al., 1987) running on an IBM XT personal computer. The contour data was transferred to an Iris Indigo workstation, where the surface renderings were generated using SYNU (Hessler et al., 1992).The amphidial neurons were reconstructed using every section; the reconstructions of cell bodies were made using only every third section. In these latter reconstructions, the positions of the cell bodies was considered important, whereas exact shape was not. The images were photographed from the workstation’s monitor.

Nomenclature The amphidial neurons in S. stercoralis were named according to the nomenclature used for these neurons in C. elegans (White et al., 1986; Sulston et al., 1988). Three letters are used per name. The first letter is A for amphid. The second letter is either S (single) for the neurons that have one dendritic process in the amphidial channel or D (double) for those with two such processes. The specialized wing cells of C. elegans are identified by W and the finger cell by F. In C. elegans, there are 12 neurons in the amphidial complex: each neuron is assigned a third letter, from A to L (Ward et al., 1975). Thus, ADF indicates a neuron with two dendritic processes, ASJ the one with a single process, and AFD the finger cell. The finger cell is a complex neuron, with dendritic processes forming a very complex cluster of 32-45 microvillilike projections; the dendritic processes of the wing cells form large flattened projections that poke into the sheath cell from the amphidial channel (Ward et al., 1975;Albert and Riddle, 1983).

RESULTS Anterior morphology When infective third-stage larvae of S. stercoralis are examined in the scanning electron microscope (Fig. l), two prominent, tripartite, lateral labia are evident. These are comprised of a prominent medial lobe and less prominent dorsal and ventral lobes. The surface of the cuticle appears wrinkled, but there are no apparent labial pores as would be expected if there were labial sensilla opening to the environment. When the dimples dorsal and ventral to the labia (evident in the scanning electron micrographs) are seen in serial sections, it is clear that they are not openings through

S. STERCORALZS AMPHIDIAL SENSORY NEUROANATOMY

283

Fig. 1. Scanning electron micrograph shows Strungylozdes stercoruthird-stage infective larva. Except for the central mouth, the only openings in the cuticle are the amphidial pores (A). Magnification,

opening in the cuticle. A small part of the outer labial neuron (OL) is found just to the left. Magnification, 40,000 x .

10,000x.

Fig. 3. Longitudinal section through an amphid of a third-stage larva. Anterior is to the left. The basal parts of five dendritic processes (P) are shown. The anterior ends of these processes do not appear in this section; a few t.ight junctions (T) at the base of the amphidial channel are evident at the right. Magnification, 20,000 X.

Zis

Fig. 2. Longitudinal section through one of the prominent lateral labia of an infcctive larva. Part ofthe lateral inner labial neuron (IL) is shown. I t forms a flattened process under the cuticle, but there is no

the cuticle. In contrast, the amphidial pores are evident laterally just posterior to the labia. Longitudinal sections in the transmission electron microscope (Fig. 2) do not show openings in the cuticle over the endings of the labial neurons or papilliform elevations, suggesting that there are no specialized sensory structures extending into the labial and adjoining cephalic cuticle. These observations indicate that the labial neurons are probably not chemoreceptors, and that they, therefore, have no significant role in the postinvasive host-recognition process (i.e., initiation of resumption of development). The ultrastructure of the labial and cephalic sensilla and their neurons in the infective third stage, as well as in other stages of S. stercorulis, will be described in subsequent publications and will not be discussed further here.

Amphids In longitudinal sections of the amphid of S. stercorulis (Fig. 31, parts of some of the dendritic processes of the amphidial neurons are seen. The processes are slightly enlarged at their anterior ends (not shown) and become thinner farther back in the amphidial channel. At the base of the amphid (Fig. 31, the neurons become larger and fan out into a much larger area. In longitudinal sections such as those shown in Figure 3, only a short length of a few of the amphidial neurons is seen, and their internal structural details are not evident. Two complete sets of transverse serial sections extending posterior from the “nose” of the worm through the nerve ring to the lateral ganglia (approximately 120 km) were

F.T. ASHTON ET AL,

284

f Figure 4

S. STERCORALJS AMPHIDIAL SENSORY NEUROANATOMY

285

examined. The amphid itself, with its ciliated dendritic process, in contrast to C. elegans, where two of the neurons processes, is approximately 9 ym long, and it is found in the each have two such processes. Using the C. elegans system first 45-50 sections. Typical sections at various levels along of nomenclature, each neuron was named by using A for the amphid are illustrated in Figure 4a-e. The first den- amphid, S for single, followed by its identifying letter. Thus, dritic process, that of neuron E, where its tip is enlarged, processes labeled E in Figure 4 and subsequent figures is nearly filling the amphidial channel (see Fig. 4f for neuro- understood to indicate neuron ASE. The lamellar cell was nal labels), is seen approximately one-fourth of a microme- named ALD, thus modifying the C. elegans nomenclature ter posterior to the amphidial pore. In the following sec- for use with S. stercoralis. tions, additional processes of dendrites appear (Fig. 4a), The lamellar cell including those of neurons C, I. and F; the number of profiles increases gradually because the processes are of In sections of the anterior end of the amphid, a system of varying lengths. Approximately 2 pm from the amphidial lamellae is found enclosed within the sheath cell (Fig. 4a). pore, 11 processes containing microtubules are seen (Fig Portions of the sheath cell form thin infoldings interdigitat4b). Approximately 6-7 ym from the anterior end of the ing with these lamellae. The lamellae are found to originate amphidial channel, the dendritic processes enlarge, spread from a dendritic process lying ventral to the amphidial out (as shown in Fig. 3), and appear to have an arrangement channel and the neurons contained within it. In subsequent of microtubule doublets, suggesting the 9 x 2 pattern (Fig. sections, the lamellar structure becomes larger and even 4c) characteristic of kinocilia (Wright, 1980). The microtu- more complex; lamellae are found on both sides of the bules end abruptly in the next micrometer along the amphidial channel, and more lamellae project ventrally amphid, and the dendritic processes now show a complex from the central shaft of the process (Fig. 4b). arrangement of lateral finlike projections running parallel In the region where the other amphidial neurons begin to to their axes (Fig. 4d). This complexity has not been show the lateral finlike projections, the central process of specifically mentioned in other nematodes, although a this lamellar cell (ALD) has merged with the other prosomewhat similar image is shown by McLaren (1976) in cesses in the amphidial channel (Fig. 4c); where the lateral Dipetalonema uiteae, a filarial nematode. There are no projections from the amphidial neurons are most extensive, similar finlike structures in C. elegans (Ward et al., 1975; the lamellar cell (ALD) has lamellae very similar in appearWare et al., 1975). ance to the projections of the other neurons (Fig. 4d). In the At the basal end of the sheath cell, approximately 9 Fm region of the tight junctions, the lamellar cell is indistinfrom the amphidial pores, all the neuronal processes form guishable in cross-sectional structure from the others, tight junctions with each other and with the sheath cell although it is slightly larger (Fig. 4e). A few lamellae are (Fig. 4e). Unlike C. elegans, where the tight junctions seen around the amphidial channel; these are posterior appear at different levels, in S. stercoralis all of the tight extensions of the lamellae seen in anterior sections. junctions are found in the same sections. There are now 13 Neuronal positional relationships neurons in the amphidial bundle: the anterior end of a very short neuron is found behind the lateral finlike projections The positions of the neurons at this level are constant of neuron E, and the dendritic process of an unusual from worm to worm. Eight additional sets of 60-65 serial neuron, which we have named the “lamellar cell,” has sections (some with sections missing or with other defects) merged with the 11others that are found in the amphidial were examined in detail, confirming the constancy of this channel. arrangement. The lamellar cell (ALD) is always found in Each of the neurons found in the amphidial channel was the same ventral-medial position, with the short neuron assigned a letter as an identifying label, as was done in C. (A) adjacent to it. The longest neuron (E) is always found elegans by Ward et al. (1975; see Materials and Methods). just lateral to these two processes in a central position in The choice of letter is explained in the Discussion section. the ventral part of the bundle. The lengths of the neurons In S. stercoralis, each neuron ends in a single dendritic also appear to be essentially constant from worm to worm.

Three-dimensionalreconstructions Fig. 4. Transverse sections at several levels through an amphid of an infective third-stage larva. a: Section approximately 1.5 km posterior to the amphidial pore. Cross sections of five neurons are found within the amphidial channel formed by the sheath cell is) at this level; one is markedly larger than the others. Lamellae of a specialized neuron, the “lamellar cell,” are evident. b: Section approximately 2.5 pm from the amphidial pore. Eleven dendritic processes are found in the amphidial channel. Complex sheets of the lamellar cell (AI,D) are contained within the sheath cell (S). c: The neurons in the amphidial channel are enlarged in this section and show a pattern of microtubules that suggest the 9 x 2 pattern characteristic of kinocilia. The lamellar cell (ALD)is less complex, and its central process lies close to the other neurons. d Section approximately 8 p,m from the amphidial pore. The amphidial dendrites show a complex pattern of finlike lateral projections. The central process of the lamellar ccll (arrow) has merged with the other neurons; at this level it has only two small lamellae. e: Section approximately 9 km from the amphidial pore. There are tight junctions between all of the neurons and between the neurons and the sheath cell. A few lamellae are evident; they are posterior extensions of lamellae seen in the anterior sections. f: Diagram of the pattern of neurons shown in e. A letter is assigned to each neuron as a label in this and subsequent figures. Magnification, 1 8 , 0 0 0 ~ .

To understand better the amphidial structure, 3-D reconstructions were made, as described in Materials and Methods. The reconstructions were viewed in two ways: as transverse images (Fig. 5a) corresponding to slices approximately six sections (or 1 pm) thick at various levels along the amphid, and in longitudinal view (Fig. 5b). Figure 5a is a stereo image of a slice 1-bm thick through the dendritic processes in the region where the lateral projections are most complex. At this level, the lamellar cell (lower right) has just two short lamellae pointing downward; the other neurons have several lateral projections each, many of which interdigitate with those of neighboring dendrites. Figure 5b is a stereo image of the reconstruction of the amphidial neurons in longitudinal view. The lamellar cell, shown as a translucent structure, “wraps” around the other neurons, which can be seen through it. Its structure cannot be seen easily in this stereo photograph of the computer reconstruction (its mottled appearance is an imaging artifact). In the comparable drawing (Fig. 61, a

286

Fig. 5. SYNU reconstructions of the amphidial dendrites. Each neuron is shown in the same color in all panels. a: Stereo image shows SYNU reconstruction of 1-pm transverse slice of the amphidial dendrites in the region where the lateral projections are most complex; compare with Figure 4d. The lamellar cell (pale blue-gray, lower right) hasjust two laniellae at this level; the other neurons have many lateral finlike projections, some of which interdigitate with those of neighbor-

F.T. ASHTON ET AL.

ing neurons. b: Stereo image of SYNU reconstruction shows the bundle of neurons within the amphidial channel in longitudinal view. The neurons are surrounded by the lamellar cell, which is shown as a translucent structure. Its complex structure cannot he seen in this image. Neuron E, in yellow, is clearly longer than the others and is enlarged at its tip.

287

S. STERCORALIS AMPHIDIAL SENSORY NEUROANATOMY

Fig. 6. Drawing of the amphidial bundle and the lamellar cell reveals its complexity and relationship to the other neurons. A section of the lamellar cell is cut out to rcvcal its lamellae and the dendritic processes it embraces.

thick slice of the lamellar neuron is removed to show its complex structure and relationship to the bundle of amphidial neurons.

Amphidial neurons and their cell bodies After leaving the tight junction region at the posterior end of the amphidial pouch, the amphidial neurons continue in a loose bundle back toward the nerve ring. The structure of this bundle is shown at four levels in Figures 7-10. A diagram of the anterior of the worm in Figure 11 shows the distances over which the neurons were traced. The profile of the lamellar cell (ALD) i s always larger than those of the other neurons, making it a convenient and easily found landmark during the tracing of neurons to their cell bodies. Approximately 4 km posterior to the tight junctions (section 64), the amphidial neurons remain in a bundle but have changed in shape (shown in Fig. 7a,b). The lamellar cell (ALD) (compare with Fig. 4f) is much larger than the others, whereas neuron A is very small. Several have mitochondria within them. Approximately 70 pm from the nose of the worm (section 390) but anterior to the nerve ring, the cell body of neuron A is found; its dendrite enters the side of the cell body (Fig. 8a,b). Similarly, in the sections following, the cell bodies of neurons B and C can be seen lying dorsal to the bundle of neurons and close to the

pharynx. The remaining amphidial neurons continue in close association to the nerve ring, approximately 95 pm from the nose of the worm, where they pass it on its lateral surface (Fig. 9a,b). They continue in close association posterior to the nerve ring where they enter their cell bodies in the lateral ganglia. Posterior to the nerve ring, approximately 110 km from the nose of the worm, the cell body of the lamellar cell, ALD, is encountered first (not shown). The connection of neuron K to its cell body, which lies dorsal to the pharynx, is found six sections posterior, in section 576 (Fig. 10a,b). At this level, the dendrites of the remaining amphidial neurons lie on the dorsal surface of the cell body of the lamellar cell (ALD). These were traced to eight additional cell bodies, four of which lie in a group lateral to the pharynx, and four others, along with neuron K, are found in a second group dorsolateral to the pharynx. A computer reconstruction of the cell bodies found anterior and posterior to the nerve ring is shown in Figure 12. This figure is approximately 47 pm long. The dendrite of the lamellar cell (ALD) is reconstructed; the others are omitted for clarity. Figure 13 shows a stereo view of a reconstruction of the 10 cell bodies found posterior to the nerve ring in S. stercoralis. The cell body on the left is that of the lamellar cell (part of its dendrite is included). Because the cell bodies on the right side of the worm are shown in this reconstruction, which is viewed from the left, the five cell bodies that lie dorsolateral to the pharynx appear in the foreground, and the other four cell bodies (along with that of the lamellar cell) that are found lateral to it are in the background. The pharynx itself is not shown. The relative positions of the cell bodies are significant, whereas the shape is not, because the tracing of every third section was used in the reconstruction. Although only 13 amphidial neurons were traced from the tips of their dendritic processes to their cell bodies, the locations of the other cell bodies in the vicinity of the nerve ring were recorded, and these structures were reconstructed. For maps of the amphidial cell bodies to be useful, the positions of these other, as yet unidentified, cell bodies had to be known. The spatial relationships between all of the cell bodies as produced from the reconstruction is shown in Figure 14b.

DISCUSSION In the infective third stage of S. stercoralis, neither the inner nor outer labial sensilla are open to the environment. The so-called cephalic sensilla also fail to communicate directly with the exterior; these sensilla are similar to those in the mouse pinworm, Syphacia obvelata (Wright, 1976). The dendrites of all these sensilla end just under the cephalic cuticle of the worm. (In the C. elegans dauer larva, however, all six of the inner labial sensilla open through pores in the cuticle; Albert and Riddle, 1983.) Because they do not open to the exterior, we did not consider either the labial sensilla or the cephalic sensilla of S. stercoralis significant in the host-finding and larval reactivation processes. The amphidial neurons are, however, likely to have a role in these processes because the amphids are open to the environment. This is supported by the observations of Bargmann et al. (1990) and of Bargmann and Horvitz (1991a,b) in C. elegans, where the amphidial neurons control entry into and exit from the dauer stage. We found

F.T. ASHTON ET AL.

288

b Fig. 7, a: Elcctron micrograph shows section 64, which is 13 pm from the nose of the worm. b: Diagram identifies the neurnns by using the letters from Figure 4f. Neuron D, the lamellar cell, is relatively large and contains four mitochondria, whereas neuron A is very small. The remaining arnphidial neurons vary in shape. Neurons B, C , K, and L each contain a mitochondrion. Other (nonamphidial) unidentified neurons are found both dorsal and ventral to t,he amphidial bundle. Magnification, 3 9 , 0 0 0 ~ .

S. STERCORALJS AMPHIDIAL SENSORY NEUROANATOMY

b Fig. 8. a: Electron micrograph shows section 390, which is 70 Km from the nose of the worm. The cell body of neuron A is shown; the remaining amphidial neurons are in a bundle just dorsal to neuron A and other neurons lie adjacent to the amphidial hundle (not identified). b Diagram identifies the amphidial neurons seen in a. Magnification, 39,000 X .

289

F.T. ASHTON ET AL.

290

b

Fig. 9. a: Electron micrograph shows section 510, which is 92 pm from the nose of the worm. Amphidial and other neurons lie in a group lateral to the nerve ring and just under the cuticle (lateral alae are evident). b: Diagram identifies the amphidial neurons in a. Magnification. 39.000 X .

13 neurons associated wit.- the amphii 11 of whit-. run in the amphidial channel and end anteriorly near the amphidial pore. The dendritic processes have different lengths that are consistent from worm to worm. The longest process, neuron E (Fig. 6a) ends in a distinct bulb close to the opening of the amphidial channel into the environment. The other 10 neurons have single processes, which are of varying, but consistent, lengths and which end in small rounded tips. As in C. elegans (Ward et al.,19751, these neurons can be identified by their individual positions at the point of entry into the base of the amphidial channel (Fig. 4fl. Just anterior to the tight junctions at the base of the amphidial channel, the dendritic processes show a complex

of lateral finlike projections running parallel to their axes (Fig. 4d). These structures are not found in C. elegans (Ward et al., 1975; Ware et al., 1975), hookworms (our own unpublished observations), or in nematode parasites of plants (Wergin and Endo, 1976; De Grisse, 1977). Possible

Fig. 10. a: Electron micrograph shows section 576, which is 105 km from the nose of thP worm. The cell body of the lamellar cell, D, is found in the lnwer left of the image; the connection of neuron K to its cell body is in the upper right. The remaining amphidial neurons, along with other neurons, lie just dorsal to the cell body of neuron D. b Diagram identifies the amphidial neurons in a. Magnification, 39,000 X .

b

F.T. ASHTON ET AL.

292

amphidial neurons

cell bodies

cell bodies

\

amphid

\

\ I I

!I

I-gp--I I

1

% ”

I

I

6 2 pm

Fig. 11. Diagram of the anterior end ofthe worm shows the location of the two groups of cell bodies on each side of the nerve ring and their approximate distances from the nose of the worm.

41 pm

I

\ \

I I

Fig. 12. SYNU rcconstructiun of the amphidial cell bodies. The dendritic process of the lamellar ccll is reconstructed; those of the other neurons were omitted for clarity. The region reconstructed is approximately 47 pm long.

Fig. 13. Stereo image shows a SYNU reconstruction of the group of amphidial cell bodies posterior to thc nerve ring. Thc lamellar ccll, including part of its dendrite, is on the lower left. Because the cell bodies on the right side of the worm are viewed from the left in this reconstruction, the five cell bodies that lie dorsolateral to the pharynx appear in the foreground, and the other four cell bodies (and that of the lamellar cell) that lie lateral to the pharynx are in the background. The pharynx itself is not shown.

functions for these lateral projections may be inferred through comparison with chemosensory neurons of insects, where the function of this class of neurons has been extensively studied (see the review by Zacharuk, 1985). Chemosensory sensilla of insects, although often more complex than those of nematodes, have a basic structure similar to that of amphids as described by Ward et al., (1975) and by De Grisse (1977). In insects, chemosensitivity of these sensilla is thought to occur through contact with chemicals in solution (i.e., in the “sensillar liquor”) in the channel within the enclosing sheath cell (Zacharuk, 1985). Various structures, including lamellae, increase the adsorp-

tive surface of the chemosensory dendrites, thus trapping stimulant molecules. In S. stercorulis, the lateral projections on the amphidial dendrites increase surface area, possibly to increase sensitivity to chemicals diffusing into the amphidial channel. The close proximity of projections from adjacent dendrites also suggests communication, either chemical or electrical, between the dendrites themselves. Lateral finlike projections are also found in the amphidial dendrites in first stage larvae of S. stercorulis. At this stage, however, the projections are much reduced in size and complexity as compared with their appearance in third-

S. STERCORALIS’ AMPHIDIAL SENSORY NEUROANATOMY

b

293

S. stercoralis

Fig. 14. Maps show the amphidial cell bodies in Caenorhabditis eleguns and Strongyloides stercoralis. a: Map shows the cell bodies in C. eleguns (modified from Sulston et al., 1988).The amphidial cell bodies

are identified; the others are left blank for clarity. The outline of the pharynx is in the background. b: Diagram shows the cell bodies in S.

stage infective larvae. It would, therefore, seem possible that, in infective larvae, they play some important, as yet unknown, role in the host-finding process. The complex dendritic process we have called the lamellar cell (ALD) has some characteristics of the finger cell (AFD) found in C. eleguns but with has some significant

stercoralis. The amphidial cell bodies are labeled according to their apparent homologous positions when cumpared with the positions of the amphidial cell bodies in C. elegan,s. The outline of the pharynx is shown.

differences. The finger cell in C. eleguns is probably the principal thermoreceptor (Perkins et al., 1986; Mori et al., 1993).Similar cells occur in many nematode species, including hookworms (El Naggar, 19871, the stomach worm of ruminants, Huemonchus contortus (our own unpublished observations), and in many plant parasites (Wergn and

294 Endo, 1976; De Grisse, 1977). As in the finger cell of C. elegans, the tip of the lamellar cell’s dendritic process is highly complex. Instead of the numerous microvillilike projections of a finger cell, it has a complex lamellar structure. Thus, in both species, the anterior terminal surface area of the dendrite is greatly increased. Its cell body, like that of the finger cell in C. elegans, is the most anterior of the amphidial cell bodies situated behind the nerve ring (Fig. 14 a,b). For these reasons, we suspect that the lamellar cell may be a homolog of the finger cell and thus will have a similar function. It would be expected that a parasite of warm-blooded organisms would have a welldeveloped thermoreceptor; the lamellar cell could well be this structure. Finger cells, however, are found dorsal to the amphidial channel (Ward et al., 1975; Wergen and Endo, 1976; De Grisse, 1977), whereas the central process of the lamellar cell lies ventral to it (Fig. 4b). In C. elegans, the dendritic process of the finger cell does not enter the amphidial channel with the other processes. Instead, the process penetrates the sheath cell directly from outside the amphidial channel where it terminates in a complex cluster of 32-45 microvilli. In S. stercoralis, the lamellar cell does enter the amphidial channel but then exits from it to poke into the sheath cell, where it forms interdigitating lamellae with complementary folds ofthe sheath cell. In this respect, the lamellar cell is more like a wing cell in C. elegans (Ward et al., 1975) than the finger cell. In the anterior of the amphid (Fig. 4a), the anterior terminal surfaces of the lamellar cell lie in a configuration somewhat similar to that of the wing cells?especially cell AWC, of C. elegans (Albert and Riddle, 1983). However, in S. stercoralis, lamellae are also found medially between the amphidial channel and the pharynx; in C. elegans, wing cell AWC lies lateral to the amphidial channel. In addition, the “wings” ofAWC in C. elegans are single-layeredstructures; the lamellar cell in S. stercoralis is complex and multilayered. The cell body of the lamellar cell lies lateral to the ventral half of the pharynx (Fig. 14b), in a position not greatly different from that ofwingcell AWC in C. elegans (Fig. 14a).Thus, the lamellar cell is similar to wing cell AWC in C. elegans in some respects. Because wing cell AWC has been shown to have a chemosenSOT function (Bargmann et al., 19931, the lamellar cell might also have this additional hnction. The thirteenth neuron found at the base of the amphidial channel appears behind the lateral processes of neuron E (Fig. 4f) and adjacent to the lamellar neuron. It is relatively short, measuring approximately 61 pm in length from its tip to its cell body (the neurons whose cell bodies are located posterior to the nerve ring are approximately twice this length). At the region of the tight junctions, its profile is always smaller than those of the other neurons.

Comparison of S. stercoralis amphidial cell bodies with those of C. elegans When the positions of the amphidial cell bodies in S. stercoralis are compared with those of the corresponding neurons in C. elegans, definite similarities are found. Figure 14a shows a diagram of the cell bodies in C. elegans (from Sulston et al., 1988). The cell bodies of the amphidial neurons; including that of the finger cell, have been identified; the identities of cell bodies of other neurons are omitted for clarity. The outline of the pharynx is shown in the background. Figure 14b shows a tracing of the recon struction made from the serial sections of S. stercoralis

F.T. ASHTON ET AL. third-stage larva. The cell body of the lamellar cell (ALD) is found in a position similar to that of the finger cell (AFD) in C. elegans. In C. elegans, there are three cell bodies, ASK, ADL, and ASI, lying dorsal to the pharynx, with cell bodies ADF and ASG just ventral and lateral to them. In S. stercoralis, there are also three cell bodies in similar dorsal positions, which we have named ASK, ASL (instead ofADL, as it has a single dendrite), and MI. Two additional cell bodies, just ventral to these, are named ASF (=ADF) and ASG . Posterior to the cell body of the lamellar cell (ALD), four amphidial cell bodies are found. Two of these have been named ASH and ASE, as their positions correspond to amphidial cell bodies with these names in C. elegans. In S. stercoralis, two additional cell bodies are found posterior to ASH and ASE; in C. elegans, only one, ASJ, is found in a somewhat more ventral position. ASJ and ASM were assigned to these cell bodies in S. stercoralis, although there is no cell body equivalent to ASM in C. elegans. Finally, the three neurons of S. stercoralrs whose cell bodies were found anterior to the nerve ring were named ASA, ASB, and ASC to complete the sequence of letters. It is not suggested that these neurons have any relationship to the wing cells in C. elegans, whose names include the letters A, B, and C, respectively (White et al., 1986). S. stercoralis and C. elegans are both rhabditid nematodes and are considered to be closely related. Holden-Dye and Walker (1994) observed that the structure of neurons in the dorsal ganglion of the intestinal roundworm, Ascaris suurn, is similar to that of the extensively mapped neurons in the dorsal ganglion of C. elegans. If neuronal structure is similar in these two more distantly related species, then function and structure are likely to be similar in more closely related species such as S. stercoralis and C. elegans. Therefore, it is reasonable to assume that neurons whose cell bodies are in similar positions might have similar functions and can be named accordingly.

Control of development by amphidial neurons Strongyloides stercoralis has a complex life cycle (Schad, 1990). Parts of this life cycle appear to be similar to certain stages in the life cycle of C. elegant$. A first-stage larva exiting a host in the feces may develop into either a free-living adult or an environmentally resistant, infective third-stage larva. It can remain dormant at this stage until a suitable host is encountered, when development can be resumed after skin-penetration. In C. elegans, there are similar developmental alternatives: a first-stage larva can develop through a series of stages to the adult stage directly, or, if environmental conditions are unsuitable, it can develop into the environmentally resistant dauer larva and remain in this stage until conditions are favorable. Bargmann and Horvitz (1991a) found that if neurons ADF, ASG. and AS1 were killed with a laser microbeam larvae of C. elegans enter the dauer stage under inappropriate conditions. If neuron ASJ is killed, larvae will not exit the dauer stage, even if conditions are favorable. The neurons we have labeled ASF, ASG, and AS1 in S. stercoralis might prove to be functional, as well as positional, homologs of their development-controlling counterparts in C. elegans. If these neurons were killed in the first-stage larvae of S. stercoralis from feces, then these larvae might all develop directly into the dauerlike infective third stage, rather than enter the indirect (heterogonic) life cycle, in which larvae develop into free-living adult worms.

S. STERCORALJS AMPHIULAL SENSORY NEUROANATOMY

Free-living adults produce eggs that give rise to infective larvae only. If neuron ASJ were killed in these larvae and the operated larvae were allowed to develop into the infective third stage, then, as suggested by studies with C. eleguns, they might not be able to resume development when exposed either in vivo or in vitro to hostlike conditions. Identification of the neuron(s) that control the infective process could provide the basis for entirely new approaches to parasite control involving interference with development at the time and place of initial contact.

ACKNOWLEDGMENTS This work was supported in part by NIH grants R 0 1 A1 22662 to G.A. Schad and RR 02483 to the IVEM and Biomedical Image Analysis Resource (Dr. L.D. Peachey, Director) at the University of Pennsylvania. This project was also supported in part by a grant from the Research Foundation of the University of Pennsylvania and by a grant from the Upjohn Corporation. We thank Xueqin Wang, of the Physics Department, for preparing the scanning electron micrograph and Susan Trammel for the drawing of the dendritic processes in the amphid. We also thank Drs. Richard Miselis and Elizabeth A. Bucher for critical comments and Dr. Frank A. Pepe for his comments and for providing space in his laboratory where the specimens were prepared for electron microscopy.

LITERATURE CITED Albert, P.S., and D.L. Riddle (1983) Developmental alterations in sensory neuroanatomy of the Caenorhabditis elegans dauer larva. J. Comp. Neurol. 198:461481. Bargmann, C.I., and H.R. Horvitz (1991a) Control of larval development by chemosensory neurons in Caenorhabditis elegans. Science 251: 12431246. Bargmann, C.I., and H.R. Horvitz (1991h) Chemosensory neurons with ovcrlapping functions: Direct chemotaxis to multiple chemicals in C. elegans. Neuron 7329-742. Rargmann, C.I., J.H. Thomas, and H.R. Horvitz (1990) Chemosensnry cell function in the behavior and dcveloprnent of Caenorhabditis alegans. Cold Spring Harbor Symp. Quant. Biol. 5.5529-538. Bargrnann, C.I., E. Hartwieg, and H.R. Horvitz (1993) Odorant-selective genes and neurons mediate olfaction in C. elegans. Cell 74:515-527. De Grisse, A.T. (1977) De ultrastructuur van het zenuwstelsel in d r kop van 22 soorten planten-parasitaire nematoden, Behorende tot 19 genera (Nematoda: Tylenchida). Dissertat.ion,Rijksuniversiteit Gent. El Naggar, H.M.S. (1987) Ultrastrudural comparison of rhabditiform (developing) and filariform (resting) larvae of the human hookworm A. duodenale. Ph.D. dissertation, University of Pennsylvania.

295

Hawdon. J., and G. Schad (1990) Serum-stimulated feeding in vitro hy infective larvae of the hookworm Ancylostoma caninurn. J. Parasit. 76:394-398. Hessler, U.. S J . Young. B.O. Carragher, M. Martone, S. Lamont, M. Whittaker. R. Mulligan, E. Masliah, J.E. Hinshaw, and M. Ellisman (1992) Programs for visualization in three-dimensional microscopy. Neurolmage 1:55-67. Holden-Dye, L., and R.J. Walker (1994) Characterization of identifiable neurones in the head ganglia of t h r parasitic nematode Ascaris m u m : A comparison with central neurones of Caenurhabdih alegans. Parasitology 108:81-87. Hotez, P., J. Hawdon, and G.A. Schad (1993) Hookworm larval infectivity, arrest and amphiparatenesis: The Caenorhabditis elegans Daf-c paradigm. Parasitol. Today 9:23-26. Jones, J.T., and I. Ap Gwynn (1991) A method for rapid fixation and dehydration of nematode tissue for electron microscopy. J. Micrnsc. I ti4:43-51. McLaren, D.J. (1976) Sense organs and their secretions. In N.A. Croll (ed): The Organization o f Nematodes. New York: Academic Press, pp. 139-161. Mori, I., H. Honda and Y. Ohshima (1993) Genes and cells important for Ihermotaxis of C. elegans. Abstracts of papers presented at the 1993 meeting on C. elegans, Meeting 319. Perkins, L.A., E.M. Hedgecock, J.N. Thomson, and J.G. Culotti (1986) Mutant sensory cilia in the nematode Caenurhabditis elegans. Dev. Biol. 117:456437. Polite, S., and M. Philip (1992) Caenorhabditis elegans as a model for parasitic nematodes: A focus on the ruticle. Parasitol. Today 8:6-12. Schad, G.A. (1990) Morphology and life history of Strongyloides stercoralzs. In D.I. Grove (ed): Strongyloidiasis: A Major Roundworm Infection of Man. London: Taylor & Francis, pp. 85-104. Sulston. J., H.R. Horvitz, and J. Kimble (1988) Cell lineage. In W.B. Wood (ed): 'The Nematode Caenorhabditis elegans. Cold Spring Harbor, NY: Cold Spring Harbor Press. Ward. S.,K.Thomson, J.G. White, and S Brenner (1975)Electron microscopical reconstruction of the anterior sensory anat,omy of the nematode Caenorhabdhs elegans. J. Comp. Neurol. 160:313-338. Ware, R.W.,D. Clark, K. Crossland, and L. Russell (1975) The nerve ring of the nematode Caenorhabditis eleguns: Sensory input and motor output. J. Comp. Neurol. 162:71-110. Wergin, W.P., and B.Y. Endo (1976) Ultrastructure of a neurosensory organ in a root-knot nematode. J. Ultrastruct. Res. 56:258-276. White, d.G., E. Southgate, J.N. Thornson, and S. Brenner. (1986) The structure of the nervous system of the nematode Caemrhabditls elegans. Phil. Trans. R. Soc. Lond. 3J4:1-340. Wright, K.A. (1976) Functional organization of the nematode's head. In N.A. Croll (ed): The Organization of Nematodes. New York: Academic Press, pp. 71-105. Wright, K.A. (1980) Nematode sense organs. In B. Zukerman (ed): Nematodes as Biological Models. New York: Academic Press, pp. 237-293. Young, S.J., S.M. Royer, P.M. Groves, and J.C. Kinnamon (1987) Threedimensional reconstruction from serial micrographs using a n ISM PC. J. Elect. Microsc. Tech. 6:207-215. Zacharuk, R.Y. (1985) Antennae and sensilla. In G.A. Kerkut and L.I. Gilbert (eds): Comprehensive Insect Physiology, Biochemistry, and Pharmacology, vol. 6: Nervous System: Sensory. Oxford: Pergammon Press, pp. 1-69.