Bacterial endosymbionts in the gills of the deep-sea wood-boring bivalves Xylophaga atlantica and Xylophaga washingtona.
In 1973 Popham and Dixon (Popham and Dickson, 1973) demonstrated the presence of endosymbiotic bacteria in the gills of wood-boring bivalves of the family Teredinidae, commonly known as shipworms. The role of these symbionts, however, was not elucidated until a decade later when it was shown that the symbiotic bacteria, when cultivated in vitro, both digest cellulose and fix atmospheric nitrogen (Waterbury et al., 1983). This novel combination of symbiont metabolic activities is now thought to be essential for the survival of teredinids on a diet composed primarily of wood, a food source that, although rich in carbon and energy, is deficient in nitrogen and cannot be digested by most animals.
Whereas teredinids are the predominant wood-boring bivalves in shallow waters (intertidal to 100 m), bivalves of the subfamily Xylophagainae (family Pholadidae) fill the same niche in the deep oceans, occurring primarily at depths from 150 m to greater than 7000 m (Turner, 1972). Like teredinids, xylophagainids bore into woody substrates, ingest the excavated wood particles, and store them in a specialized caecum formed by an outpocketing of the stomach (Purchon, 1941; Turner, 1973). The role of wood in the diet of Xylophaga has not, however, been determined experimentally. On the basis of morphOlogical and ecological arguments, it has been proposed that Xylophagainids derive (1) significant (Purchon, 1941; Turner, 1973), (2) little (Potts, 1923), or (3) no (Yonge, 1937, 1938; Yonge and Thompson, 1976) nutrition from the ingestion of wood.
The utilization of wood by xylophagainids is a subject of considerable ecological interest since wood is abundant in many areas of the deep sea (Wolff, 1979) and the opportunistic Xylophagainae are among the most common colonizers of wood at these depths (Turner, 1973). Hence the deep-sea Xylophagainae could perform a role analogous to that of termites in terrestrial habitats and shipworms in shallow marine habitats by converting the refractory cellulosic carbon deposited in wood and other plant remains to a more readily available form (Turner, 1973). Both termites (Kane, 1997) and shipworms (Waterbury et al., 1983) utilize cellulolytic and nitrogen-fixing symbionts to survive on wood as the sole food source. In this report we present evidence that bacterial endosymbionts are present in the gill tissue of two species of Xylophaga, and that these bacteria appear morphologically similar to the gill endosymbionts previously identified in shipworms.
Materials and Methods
Specimens of X. washingtona were collected in pine boards submerged for 2-3 months in Monterey Bay (depth 61 m, coordinates 36 [degrees] 39.8 [minutes] N, 121 [degrees] 52.88 [minutes] W; courtesy of Dr. E.C. Haderlie, Naval Postgraduate School) or in Scripps canyon off the San Diego coast (depth 274 m, coordinates 32 [degrees] 31 [minutes] N, 117 [degrees] 16.5 [minutes] W). Specimens of X. atlantica were collected from oak boards (1[inch] x 2[inches] x 2[feet] lobster-trap skids) submerged for about 1 year at 80-100 m depth 12 miles off the coast of SW Harbor, Maine. Animals were kept alive in chilled seawater tanks until they were removed from the wood and prepared for microscopy (less than 2 weeks). Animals with an average valve diameter of 3-4 mm were dissected and fixed for 1-1.5 h in 3% glutaraldehyde buffered with 0.1 M cacodylate/HCl (pH 7.3) and 0.4 M NaCl or 3% glutaraldehyde buffered with 0.1 M sodium phosphate (pH 7.4), 3% NaCl, and 4.5% sucrose as in Eckelbarger et al. (1990). After fixation the specimens were post-fixed in 1% osmium tetroxide and dehydrated through a graded ethanol series. Specimens for transmission electron microscopy (TEM) were then transferred to propylene oxide and embedded in Epon/Araldite for sectioning. Specimens for scanning electron microscopy (SEM) were fractured under liquid nitrogen after ethanol dehydration and then were critical-point dried from C[O.sub.2] and sputter coated with gold. A Phillips CM 10 transmission electron microscope and an AMR 1000 scanning electron microscope were used to examine samples.
Both Xylophaga atlantica and Xylophaga washingtona were used for TEM of the gills and light microscopy of the digestive tract. Similar results were observed with both species. X. atlantica was used for SEM of the gills, and X. washingtona was used for TEM of the digestive tract.
Gross morphology of the ctenidium
The gross morphology of the gills of Xylophaga has been described in detail by Purchon (1941) and will be briefly summarized here. The ctenidia (gills) of Xylophagainae, like those of the Teredinidae, are formed by a single (outer) demibranch on either side of the visceral mass [ILLUSTRATION FOR FIGURE 1 OMITTED]. The marginal groove is absent, there is no ciliary sorting mechanism, and the labial palps are greatly reduced. These features are shared by most teredinids but are divergent from the filter-feeding gill morphology of other lamellibranchs (Purchon, 1941). The ctenidia are connected to the visceral mass dorsally at the efferent branchial vein and ventrally at the afferent branchial vein, forming an arc that separates the mantle cavity into suprabranchial and infrabranchial chambers. Each filament is composed of two distinct regions: a narrow frontal region (fr) and a broad, flattened abfrontal region (ar). As is typical of lamellibranch gills, the frontal region is composed of frontal, laterofrontal, and lateral ciliated cells; a central blood vessel; and supporting connective tissue. The abfrontal region, however, is distinctive, consisting of an extensively developed interlamellar junction that forms a crescent-like sheet bridging the arc of the filament [ILLUSTRATION FOR FIGURE 1 OMITTED]. The filaments are joined laterally by interfilamentar cellular junctions in the frontal region. Whereas cellular interfilamentar junctions are absent in the abfrontal region, filaments appear to be connected by lateral ciliary junctions in this region [ILLUSTRATION FOR FIGURE 2E OMITTED].
Identification and localization of intracellular bacteria in the ctenidium
In the Xylophagainae, as in other symbiont-bearing bivalves, the frontal region of each gill filament is devoid of symbionts. The most conspicuous cytoplasmic components of the cells of the frontal region include mitochondria, Golgi, endoplasmic reticulum, ciliary structures, centrioles, and glycogen granules. Bacterial symbionts are, however, abundant in the abfrontal region of the ctenidium [ILLUSTRATION FOR FIGURE 2D-F OMITTED] AND [ILLUSTRATION FOR FIGURE 3 OMITTED]. This region forms a broad, flattened plate composed of two closely appressed single-layered sheets of epithelial cells surrounding a narrow blood sinus.
The epithelium of the abfrontal region consists of two recognizable cell types here referred to as bacteriocytes and intercalary cells in keeping with terminology in current use for other bivalve symbioses (Fisher, 1990; Frenkiel et al., 1996; Gros et al., 1996). Bacterial endosymbionts are absent from the intercalary cells but dominate the cytoplasm of the bacteriocytes. The symbionts are distributed in clusters throughout the bacteriocytes, with mitochondria infrequently interspersed in the cytoplasm between clusters. Large membrane-bound inclusions, possibly lysosomes or lysosomal residual bodies, also occur in the bacteriocyte cytoplasm [ILLUSTRATION FOR FIGURE 3 OMITTED]. These inclusions frequently display a biphasic appearance with a granular low-density region and a slightly denser reticulate region that appears to consist of whorled membranes. Some inclusions contain bodies suggestive of partially degraded bacteria. Bacteriocytes and intercalary cells are uniformly distributed in the epithelium of the abfrontal region with about equal frequency. The apical surfaces of both cell types are densely covered with microvilli at points that directly contact the external environment [ILLUSTRATION FOR FIGURE 3C OMITTED].
The bacteriocytes are typically spherical or cylindrical. Their broad basal surfaces compose much of the internal lining of the blood sinus. Intercalary cells, on the other hand, are typically narrow at their basal ends, expanding to a broad and irregular apical surface. Thin sheet-like projections of the apical end of the intercalary cells extend over the apical surfaces of adjacent bacteriocytes, apparently shielding most but not all of the surface of the bacteriocytes from direct contact with the external environment. The outer surface of the abfrontal epithelium is elaborated into a series of larger rounded hummocks (formed by the spherical outline of the bacteriocytes) and smaller papillae (formed by the irregular surfaces of intercalary cells) surrounded by many deep folds and pits [ILLUSTRATION FOR FIGURE 2F OMITTED] AND [ILLUSTRATION FOR FIGURE 3A OMITTED]. The inner blood-facing surface of this epithelium also contains numerous invaginations [ILLUSTRATION FOR FIGURE 3 OMITTED]. These elaborations greatly increase the surface areas that are exposed to blood internally and to seawater externally in the symbiont-containing abfrontal zone.
The endosymbionts are straight to gently curved rods that display a cell-wall morphology typical of gram-negative bacteria and range from 0.4 to 0.7 [[micro]meter] in width and up to 5.0 [[micro]meter] in length [ILLUSTRATION FOR FIGURE 3 OMITTED]. No conspicuous internal structures, such as the sulfur granules seen in thiotrophic symbionts or the stacked internal membranes seen in methanotrophic symbionts of other bivalve families (Fisher, 1990), are observed in the symbionts of Xylophaga.
The symbionts are not in direct contact with the cytoplasm but are contained within membrane-bound vesicles ranging from 10 to 20 [[micro]meter] in diameter. A single vesicle may contain many bacteria. The orientation of symbionts within vesicles is not random. Most symbionts are oriented with their long axes perpendicular to the lateral surface of the filament. Few symbionts are oriented with their long axes parallel to the lateral surface and of those, fewer still align with the frontal-abfrontal axis of the filament. This distribution of orientations is readily observable in both SEM and TEM images.
Extensive examination of TEM images failed to provide evidence of connections between the symbiont-containing vesicles and the external environment. In all sections examined, the vesicular membrane appears distinct from and unconnected to the plasma membrane. Similarly, examination of the exterior surfaces by SEM revealed no evidence of the kind of ducts or openings to the bacteria-containing vesicles that are seen in the symbiont-containing light organs of some luminous fish (Haygood, 1993). Although deep pits and infoldings were observed on the lateral surfaces of the filaments in SEM images, TEM images showed that these are not continuous with the symbiont-containing vesicles [ILLUSTRATION FOR FIGURE 3C OMITTED]. Hence, at minimum, the bacterial symbionts are separated from the external environment by the plasma membrane, cytoplasm, and vesicular membrane of the host.
Wood and flora of the digestive tract
Xylophaga has a digestive system similar to that described for shipworms, including a large caecum that is typically distended with wood shavings [ILLUSTRATION FOR FIGURE 4 OMITTED]. The dimensions of the shavings (width and height) match the contours of the dentition on the shell, consistent with the ingestion of wood excavated during burrowing. Wood was the predominant component of the gut contents identifiable at the level of light microscopy. The caecum was the most densely packed space, while the stomach contained dispersed wood fibers. Fibrous material was found in proximity with the crystalline style and at low density in the style sac. Examination of serial thick sections of the guts of several animals failed to reveal any conspicuous community of microorganisms resembling those found in wood-eating insects. Electron microscopy confirmed this observation and further indicated that bacteria are absent from cells of the ciliated epithelium lining the gut and occur only sparsely in the digestive spaces [ILLUSTRATION FOR FIGURE 4D-E OMITTED].
The discovery of bacterial endosymbiont populations in a number of marine invertebrate species has been the key to understanding the remarkable ability of these invertebrates to survive on unusual nutrient sources. For example, the identification of cellulolytic, nitrogen-fixing bacteria in the gills of the wood-boring teredinid clams (shipworms) pointed toward an explanation of how these bivalves are able to thrive with wood as their primary food source. This capability, which is quite rare among higher animals, has led to the great success of the Teredinidae as colonizers of wood in coastal marine environments. In the deep oceans, the dominant colonizers of wood and other plant materials are the Xylophagainae (family Pholadidae). Similarities in the gut anatomy and wood-boring habits of the Teredinidae and the Xylophagainae led us to examine whether species of Xylophaga maintain similar symbiotic associations. In this report we present morphological evidence for the existence of bacterial endosymbionts in the gill tissue of two of these deep-sea wood-boring bivalves - Xylophaga atlantica and Xylophaga washingtona.
In many respects, the gill of Xylophaga closely resembles the symbiont-containing gills of other bivalves including the Lucinidae, Vesicomyidae, Modiolinae, and Solemyidae (Distel and Felbeck, 1987; Fisher 1990; Frenkiel et al., 1996; Gros et al., 1996). The gill of Xylophaga is composed of two regions: a heavily ciliated, symbiont-free frontal region that is like the typical lamellibranch gill structure, and an abfrontal region that appears to be specifically adapted to harbor symbiotic bacteria. The abfrontal region, an extension of the interfilamentar junction, is much broader than the frontal region and composes the bulk of the mass of the gill. The symbionts are completely surrounded by a vacuolar membrane of host origin and lack contact with the external environment. In addition to harboring symbionts, the bacteriocytes contain large membranous structures resembling lysosomes in various stages of development. In some instances, these lysosome-like compartments contain bodies that resemble partially degraded bacteria, which suggests that lysosomal digestion is a mechanism for regulation of the bacterial population.
The symbiont-containing tissue in Xylophaga also resembles that of the teredinids. Although teredinid gills are highly modified, the symbionts are found in an anatomically analogous region (Distel et al., 1991). This region is referred to as the gland of Deshayes (Sigerfoos, 1908) in teredinids and corresponds to the interlamellar junction of the fused right and left demibranchs (Turner, 1966; Distel et al., 1991). Bacteriocytes in this tissue contain symbionts that resemble the symbionts of Xylophaga; both are gram-negative rods averaging about 0.5 [[micro]meter] in width and up to 5 [[micro]meter] in length, and both lack internal membranes or other conspicuous internal or external structures.
It is striking that xylophagainids and teredinids not only harbor symbionts that are similar in appearance and location, but that they also share similar adaptations to their woody habitats. Both are obligatory wood borers that excavate burrows by using their shells as rasps, ingest the excavated wood shavings, and store the wood particles in a large caecum prior to passing them through the stomach and gut. Purchon (1941) argued that several features common to the xylophagainid and teredinid digestive systems indicate a departure from reliance on filter feeding and represent specific adaptations to the digestion of wood. These include the lack of a marginal food groove, limited ciliary sorting mechanisms in the ctenidia, reduced labial palps and crystalline style sac, and presence of a wood-storing caecum with ciliated grooves capable of providing a steady stream of wood particles to the stomach.
It is now well established that wood is a primary constituent of the teredinid diet (Gallager et al., 1981), although nutrients may also be obtained from suspension feeding (Mann and Scott, 1985). In fact, at least one shipworm species has been shown to be capable of sustaining normal growth and reproduction with wood as the sole source of particulate nutrients (Gallager et al., 1981). The shipworm symbiont has been cultivated in vitro and shown to fix atmospheric nitrogen and secrete cellulolytic, xylanolytic, and proteolytic enzymes (Greene and Freer, 1986; Greene et al., 1988; Greene, 1989, Greene and De Wispeleare, pers. comm.). Although these symbiont activities are appropriate for a wood-based diet and are thought to be critical for the shipworm's success in colonizing woody substrates, no direct evidence exists for their participation in the nutrition of the shipworm hosts.
Since the Xylophaga symbiont has not yet been cultivated, the question of whether its metabolic capabilities are similar to those of the shipworm symbiont is unresolved. Hence, it is possible that the two bacteria play different roles in their respective symbioses. For example, the bacterial symbionts in Xylophaga may be nitrogen fixing but not cellulolytic, or they may contribute essential nutrients that are absent from wood. On the other hand, phylogenetic analyses based on 16S rRNA sequences indicate that the Xylophaga symbiont is the closest relative of the shipworm isolates identified to date (Distel and Roberts, 1996). That finding increases the likelihood that these two symbionts play similar physiological roles.
Although the presence of similar bacteria in the same regions of the gills of the Teredinidae and the Xylophagainae does not demonstrate a common function for the two symbioses, the additional features they have in common - their wood-boring habits and parallel anatomical modifications of the digestive tract - lend support to this hypothesis. To our knowledge, no animal species has yet been demonstrated to utilize wood as a primary nutrient source without the aid of symbiotic microorganisms; thus if wood is the primary dietary constituent of Xylophaga, a nutritional role for its symbionts is strongly implicated. If wood is not a primary food source for the xylophagainids (as it is for teredinids), it is difficult to explain the function of the caecum, account for the fact that wood is the principal constituent of the caecal and gut contents, or identify the diet of the xylophagainids. The scarcity of alternative food sources in the deep sea, as well as the notable paucity of bacteria, phytoplankton, and other microorganisms in the gut of Xylophaga, seems to preclude the possibilities that Xylophaga subsists by filter feeding, by grazing on wood-associated microorganisms, or by utilizing an extracellular symbiotic gut microfauna as do termites and ruminants. Nonetheless, the possibility cannot be ruled out that Xylophaga (or the teredinids) produces sufficiently active endogenous cellulases to facilitate wood digestion, as has been suggested for some termites (Breznak and Brune, 1994).
To determine whether the Xylophaga symbiosis is cellulolytic, it is necessary to identify the activity of cellulolytic or nitrogen-fixing enzymes in the symbionts, demonstrate the depletion of cellulose in fecal material, and analyze the nutritional utilization of carbon derived from cellulose. These properties have been demonstrated for the shipworm symbiosis, although it is still not clear how the cellulolytic enzymes produced by bacterial endosymbionts in the gill might be transported to the gut where wood digestion must occur. In the teredinids, it has been proposed that a duct in the afferent branchial vein connects the gills to the esophagus (Saraswathy, 1971). Such a duct could serve as a conduit for cellulolytic enzymes; however, the complete vessel was not observed in all species examined, and its existence has not been independently confirmed. Our results indicate that the symbionts of Xylophaga (this study) and those of teredinids (unpubl. obs.; Trytek and Allen (1980)) are truly intracellular. Consequently, the plausibility of a symbiotic mechanism depends on the demonstration of a pathway whereby enzymes from an intracellular bacterial population are ultimately transported to the lumen of the gut to participate in cellulose degradation.
We thank Dr. Ruth D. Turner and Dr. George Somero for helpful discussion and encouragement; Dr. Kevin Eckelbarger, Dr. Nick Holland, and Kelly Edwards for expertise in electron microscopy of marine invertebrates; Wendy Morrill for technical assistance; and Jack Merrill, Ron McConnaughey, and Dr. E. C. Haderlie for aid in collecting specimens. This work was supported by NSF EPSCoR Grant # EHR 91-08766, awards from the Regular Faculty Research and Summer Faculty Research Funds of the University of Maine, Orono, and a research award from the Scripps Industrial Associates.
Breznak, J. A., and A. Brune. 1994. Role of microorganisms in the digestion of lignocellulose by termites. Annu. Rev. Entomol. 39: 453-487.
Distel, D. L., and H. Felbeck. 1987. Endosymbiosis in the lucinid clams Lucinoma aequizonata, Lucinoma annulata and Lucina floridana: a reexamination of the functional morphology of the gills as bacteria-bearing organs. Mar. Biol. 96: 79-86.
Distel, D. L., and S. J. Roberts. 1996. Bacterial endosymbionts in the gills of the deep-sea, wood-boring pholad clams Xylophaga atlantica and Xylophaga washingtona. P. 350 in Abstracts: American Society for Microbiology 96th General Meeting, New Orleans, LA, American Society for Microbiology, 1325 Massachusetts Ave., N.W., Washington, DC.
Distel, D. L., E. F. Delong, and J. B. Waterbury. 1991. Phylogenetic characterization and in situ localization of the bacterial symbiont of shipworms (Teredinidae: Bivalvia) by using 16S rRNA sequence analysis and oligonucleotide probe hybridization. Appl. Environ. Microbiol. 57: 2376-2382.
Eckelbarger, K., B. Bieler, and P. Mikkelsen. 1990. Ultrastructure of sperm development and mature sperm morphology in three species of commensal bivalves (Mollusca: Galeommatoidea). J. Morphol. 205: 63-75.
Fisher, C. R. 1990. Chemoautotrophic and methanotrophic symbioses in marine invertebrates. Rev. Aquat. Sci. 2: 399-436.
Frenkiel, L., O. Gros, and M. Moueza. 1996. Gill structure in Lucina pectinata (Bivalvia: Lucinidae) with reference to hemoglobin in bivalves with symbiotic sulphur-oxidizing bacteria. Mar. Biol. 125: 511-524.
Gallager, S. M., R. D. Turner, and C. J. Berg. 1981. Physiological aspects of wood consumption, growth, and reproduction in the shipworm Lyrodus pedicellatus Quatrefages. J. Exp. Mar. Biol. Ecol. 52: 63-77.
Greene, R. V. 1989. A novel, symbiotic bacterium isolated from marine shipworm secretes proteolytic activity. Curr. Microbiol. 19: 353-356.
Greene, R.V., and S.N. Freer. 1986. Growth characteristics of a novel nitrogen-fixing cellulolytic bacterium. Appl. Environ. Microbiol. 52(Nov.): 982-986.
Greene, R. V., H. L. Griffin, and S. N. Freer. 1988. Purification and characterization of an extracellular endoglucanase from the marine shipworm bacterium. Arch. Biochem. Biophys. 1(15): 334-341.
Gros, O., L. Frenkiel, and M. Moueza. 1996. Gill ultrastructure and symbiotic bacteria in the tropical lucinid, Linga pensylvanica (Linne). Symbiosis 20: 259-280.
Haygood, M. G. 1993. Light organ symbioses in fishes. Crit. Rev. Microbiol. 19(4): 191-216.
Kane, M.D. 1997. Microbial Fermentation in Insect Guts. Chap. 8, pp. 231-265 in Gastrointestinal Fermentation and Ecosystems, Vol. 1, R. 1. Mackie and B. A. White, eds. Chapman and Hall, New York.
Mann, R., and M. G. Scott. 1985. Growth, morphometry, and biochemical composition of the wood boring molluscs Teredo navalis L., Bankia gouldi (Bartsch), and Nototerendo knoxi (Bartsch). J. Exp. Mar. Biol. Ecol. 85: 229-251.
Popham, J. D., and M. R. Dickson. 1973. Bacterial associations in the teredo Bankia australis (Lamellibranchia, Mollusca). Mar. Biol. 19: 338-340.
Potts, F. A. 1923. The structure and function of the liver of Teredo, the shipworm. Proc. Camb. Phil. Soc. 1: 1-17.
Purchon, R. D. 1941. On the biology and relationships of the lamellibranch Xylophaga dorsalis (Turton). J. Mar. Biol. Assoc. U.K. 25(1): 1-39.
Saraswathy, M. 1971. Observations on the structure of the shipworms, Nausitora hedleyi, Teredo furcifera and Teredora pricesae (Bivalvia: Teredinidae). Trans. Roy. Soc. Edinb. 68(14): 508-562.
Sigerfoos, C. P. 1908. Natural history, organization and late development of the Teredinidae or shipworms. Bull. Bur. Fish. 37: 191-231.
Trytek, R. E., and W. V. Allen. 1980. Synthesis of essential amino acids by bacterial symbionts in the gills of the shipworm Bankia setacea (Tryon). Comp. Biochem. Physiol. 67A: 419-427.
Turner, R. D. 1966. A Survey and Illustrated Catalogue of the Teredinidae (Mollusca: Bivalvia). The Museum of Comparative Zoology, Harvard University, Cambridge, MA.
Turner, R. D. 1972. A new genus of deep water wood-boring bivalve (Mollusca, Pholadidae, Xylophagainae). Basteria 36(2-5): 98-104.
Turner, R. D. 1973. Wood-boring bivalves, opportunistic species in the deep sea. Science 180: 1377-1379.
Waterbury, J. B., C. B. Calloway, and R. D. Turner. 1983. A cellulolytic-nitrogen fixing bacterium cultured from the Gland of Deshayes in shipworms (Bivalvia: Teredinidae). Science 221: 1401-1403.
Wolff, T. 1979. Macrofaunal utilization of plant remains in the deep sea. Sarsia 1(2): 117-136.
Yonge, C. M. 1937. Evolution and adaptation in the digestive system of metazoa. Biol. Rev. 12: 87-115.
Yonge, C. M. 1938. Recent work on the digestion of cellulose and chitin by invertebrates. Sci. Prog. 32: 638-47.
Yonge, C. M., and T. E. Thompson. 1976. Borers in rock and timber. Chap. 16, pp. 205-219 in Living Marine Molluscs. William Collins Sons, London.