Fossil traces of the bone-eating worm Osedax in early Oligocene whale bones

Abstract

Osedax is a recently discovered group of siboglinid annelids that consume bones on the seafloor and whose evolutionary origins have been linked with Cretaceous marine reptiles or to the post-Cretaceous rise of whales. Here we present whale bones from early Oligocene bathyal sediments exposed in Washington State, which show traces similar to those made by Osedax today. The geologic age of these trace fossils (∼30 million years) coincides with the first major radiation of whales, consistent with the hypothesis of an evolutionary link between Osedax and its main food source, although older fossils should certainly be studied. Osedax has been destroying bones for most of the evolutionary history of whales and the possible significance of this “Osedax effect” in relation to the quality and quantity of their fossils is only now recognized.

The deep sea has the least explored biodiversity (1, 2), and the scarcity of food in the abyss has resulted in a range of evolutionary novelties (3–5). A recent discovery in this field is the annelid genus Osedax that lives and feeds exclusively on bones on the seafloor (6). Age estimates using molecular clocks suggest either an Eocene to Oligocene origin of Osedax, coincident with the rise of whales (6, 7), or a Cretaceous origin (7), depending on the rate used, but these estimates have not yet been corroborated by fossil evidence. Osedax belongs to the family Siboglinidae that includes the large tube worms living around deep-sea hydrothermal vents and cold seeps (6). Whereas other siboglinids live in symbiosis with chemoautotrophic bacteria, Osedax has symbionts that are heterotrophic γ-proteobacteria consuming mainly collagen and/or lipids (8). The symbionts are housed mainly in tissue that forms a “root system” extending into the bone. The action of the roots and associated bacteria results in the destruction of the bone interior. The roots are connected to the main body of the worm that emerges from the bone via a circular hole on the bone surface (6, 9). Such holes and excavations in fossil bones can arguably be used to infer the presence of Osedax in the geologic past. Here we report Oligocene whale bones that show such traces.

Results

Traces resembling those left by Osedax in whale bones today were found in two early Oligocene whales from bathyal sediments of the Pysht Formation in northwestern Washington State (Fig. 1). The whale fossils were preserved within hard carbonate concretions from rock outcrops on the modern beach terrace. The whales were small, toothed mysticetes with a body length estimated to not exceed 4 m. One specimen (USNM 539939) is the posterodorsal part of a skull that includes a right dentary, a periotic, a bulla, some teeth, and other fragments in addition to six small shark teeth (?Somniosus sp.). Boreholes are on the lateral surface of the dentary and on two rib fragments (Fig. 1 A and C). Some of the bones preserve marks left by the teeth of scavenging sharks. The ventral portion of the skull was corroded away before being fossilized, likely due in large part to the activities of Osedax. The other specimen (UWBM 91837) is the ventral part of a skull with the earbones (both periotics and bullae) in place. Boreholes are on the posterior face of the right jugular process. The dorsal side of this skull had corroded away before being fossilized, likely due to Osedax.

Fig. 1.

Fossil and recent whale bones bored by Osedax. (A) Rib fragment of an early Oligocene whale with Osedax borings (specimen no. USNM 539939). View is of surface with circular boreholes; those shown in the micro-CT–based visualization in E are numbered; arrows delimit reconstructed section in E. tb, remnants of trabecular bone. (B) View of modern whale bone colonized by O. rubiplumus for comparison. (C) Lateral view of the same fossil rib fragment as in A, showing the excavated interior. (D) Modern whale bone colonized by O. roseus, with the surface layer of the bone partially removed to show roots in the destroyed interior. Note lobe-like roots as well as straight tube-like roots. (E) Micro-CT–based rendering of the trace fossils. Bone material is in transparent blue, and boreholes and cavities are in yellow. (F) Piece of gray-whale bone sliced open with a scalpel exposing the ovisacs and roots of several O. rubiplumus specimens. The specimen emerging to the left had colonized an already exposed interior bone matrix and created a large cavity and roots extending parallel with the bone surface. Other O. rubiplumus that colonized the intact surface of the bone and penetrated the interior matrix, leaving several cavities, are visible.

The traces that we attribute to Osedax start as boreholes on the bone's surface and lead to cavities inside (Figs. 1E and 2). The boreholes are circular to slightly irregular in outline with a diameter ranging from 0.10 to 0.45 mm, which is consistent with the trunk diameter of smaller species of Osedax. Some smaller surface pits, with a diameter of up to 0.20 mm, that did not lead into cavities underneath were also seen. The figured rib fragment has an average of 8 boreholes/1 cm² surface area that lead into cavities beneath the bone; the distance from one successful borehole to its nearest neighbor ranges from 0.7 to 4.4 mm. On the dentary of the same specimen (USNM 539938) are 7.5–11 borings/cm²; on the other specimen (UWBM 91837) are up to 15.5 holes/cm². The only bones that showed no borings are the earbones (tympanic bullae and periotics); these bones are densely ossified with distinctively hard, almost glassy, osteosclerotic bone (10) and may have been too hard for Osedax to penetrate or may have lacked enough nutrients to sustain the worms.

Fig. 2.

Cavity morphology in early Oligocene rib fragment (USNM 539939). (A) Reconstructed image of a CT scan through holes 1 and 3 in Fig. 1 A and E. Note destroyed trabecular bone just below the left borehole. (B) Micro-CT–based rendering of a cavity in solid bone; only a single borehole leads into it, and it is assumed to have been excavated by a single individual of Osedax. Note the difference between the long and thin vascular canals and the thicker and more irregularly shaped root lobes. (C) Reconstructed image of a CT scan of the same cavity as in B; the three small, elliptical holes in the Upper half are vascular canals. (D) Micro-CT–based rendering of another cavity in solid bone. b, borehole on bone's surface; rl, root lobe; sb, solid bone; tb, trabecular bone; vc, vascular canal of the bone.

The cavities inside the bone show different shapes depending on the structure of the bone, as visualized in the micro-CT–based renderings (Figs. 1E, 2). All numbered boreholes on Fig. 1A lead into a single cavity that occupies most of the interior of the figured specimen. Approximately 10 boreholes/1 cm² of cavity base area lead into this cavity, and it shows various degrees of destruction of the bone. Below borehole number 1 in Fig. 1A only the area just beneath the surface is destroyed, and the trabecular (spongy) bone below appears to be intact (Fig. 2A, Left). Below boreholes 6 and 7 in Fig. 1A the interior of the bone is completely excavated; only the floor of the bone still shows remnants of the original trabecular structure (Far Right in Fig. 1 A and C). These remnants of trabecular bone on the margins of this cavity with multiple boreholes suggest that the original bone had mostly trabecular structure. Cavities in mostly solid bone are smaller and consist of a tube leading from the bone's surface to its interior where the cavity branches into several lobe-like extensions and thin elongate tubes approximately parallel to the bone's surface. The spaces leading from the bone surface are cylindrical or globular in shape, have approximately the same diameter as the hole in the bone's surface or are slightly wider, and are about 0.2–0.3 mm high, consistent with the base of an Osedax trunk and the ovisac region. The lobe-like extensions are interpreted here as having housed the branched root system of Osedax, whereas some of the elongate tube-like cavities could also represent relatively straight extensions of root tissue that are seen in some Osedax species (Fig. 1D). Other elongate tubes are better interpreted as part of the vascular system of the bone itself. The cavities in solid bone investigated here (n = 3) reach a maximum depth of 1 mm below the bone's surface and a maximum lateral extension of 1.7 mm (excluding the long tubes that we interpret as vascular canals of the bone), and they are here considered to reflect the size and shape of the Oligocene Osedax root system.

Discussion

The traces in Oligocene whale bones reported here resemble those produced by Osedax in modern whale bones (Fig. 1). Bone diseases do not result in circular holes on the bone's surface (11), and the few other deep-water organisms boring into hard substrates produce borings of distinctly different shape. Xylophagain bivalves typically bore into wood, but have also been reported to have bored into gutta-percha sheaths of deep-sea cables (12), and their wood-boring behavior is known as far back as the late Cretaceous (13). Borings of xylophagains are elongate-conical to pear-shaped and thus unlike the borings reported here. Sipunculid worms are also known to bore into hard substrates, mainly carbonate, and have a broad depth range (14). Sipunculid borings are elongate, narrow, nonbranching, and straight or winding tubes (15) and thus very different from the borings of Osedax. A sipunculan associated with whale bones in the deep sea was found nestling in crevices and between the attached modiolid mussels, but it was not reported to have bored into the bone (16). Clionid demosponges produce similarly sized boreholes, although they tend to be more irregularly shaped than those of Osedax, and their excavations consist of numerous extremely thin and branching filaments (17), unlike those reported here. Furthermore, clionids are known to bore chemically only into calcium carbonate (18) and are thus unlikely to penetrate bone material. Studies on six whale carcasses from 380 to 2890 m depth from Monterey Bay in California have examined many corroded modern whale bones and cataloged the fauna, but no bone-boring animals other than Osedax have been located (7, 19).

The boreholes and cavities in our early Oligocene whale bones are at the smaller end of the size range of boreholes, trunks, and roots of living Osedax (Table 1). However, it is difficult to deduce the possible phylogenetic relationships of the Oligocene species to modern ones from these size data because species with very different trunk sizes are genetically very closely related (e.g., O. mucofloris and O. japonicus or O. rubiplumus and O. roseus (7). The roots of most Osedax species are robust lobate (Table 1) although in some species it can also be branching filiform (e.g., Osedax roseus; compare images in refs. 7 and 20 and our Fig. 1D). The shape of the roots to some extent may result from the structure of the bone that it penetrates. Thus whether the differently shaped cavities in our Oligocene bones were produced by different species of Osedax or by a single species that produced different cavities depending on the bone structure is uncertain. Modern whale carcasses can also show a succession of colonization by different Osedax species, and different species can colonize the same bone without evidence for spatial segregation (19). The smaller pits that were observed but did not lead into cavities beneath the bone could be interpreted as failed Osedax borings, but there is little evidence from the activities of present day Osedax to support this hypothesis. These small pits could be the result of some other process and further investigation is required.

Table 1.

Size of trunk and boreholes and shape of roots in modern Osedax and the traces reported here

Osedax species	Trunk diameter (mm)*	Borehole diameter (mm)	Root size (mm)	Root shape	Sources including their figure numbers
O. frankpressi	0.9–1.6		7.5	Robust lobate	Fig. 2 c and d in ref. 6
O. japonicus	1.1–2.1		4.3	Robust lobate	Fig. 3 a and I in ref. 9
O. mucofloris	0.4–0.5		10 (2, 3†)	Robust lobate	Fig. 1d in ref. 30
O. roseus	0.5–0.9	0.1–1.0	17.5	Robust lobate or branching filiform	Fig. 1e in ref. 7; figs. 3 c and d and 4 a and b in ref. 20
O. rubiplumus	1.9–2.0	0.2–2.0‡	38	Robust lobate or branching filiform	Fig. 1C in ref. 6; fig. 1f in ref 7
Osedax “nude palp C”	0.5				Fig. 1i in ref. 7,
Osedax “nude palp D”	0.6				Fig. 1j in ref. 7
Osedax “orange collar”	0.7	0.1–1.0‡			Fig. 1a in ref. 7; fig. 3I in ref. 19
Osedax “spiral”	1.2	NA, lives in sediment	10	Robust lobate	Fig. 3F in ref. 19; fig. 1g in ref. 7
Osedax “white collar”	1				Fig. 1c in ref. 7
Osedax “yellow collar”	0.6–0.8	0.1–1.0‡	7.5	Branching lobate	Fig. 3J in ref. 19
Osedax “yellow patch”	0.5		15+	Robust lobate	Fig. 1h in ref. 7
Fossil traces		0.1–0.45	1.7	Branching lobate	This study

NA, not applicable.

*Published measurements or measured at base of trunk on published images.

^†Reported size of excavation within bone.

^‡G. W. Rouse (personal observations).

Our 30-million-year-old trace fossils fall within the range of the molecular age estimates (6, 7) for Osedax and coincide with the first major diversification of whales (21). Although it seems plausible that the diversification of whales facilitated the diversification of Osedax, it does not necessarily relate to its origin. When using the slow substitution rates found in some deep-sea annelids rather than those of shallow-water invertebrates, molecular age estimates suggest a Cretaceous origin of Osedax (7). In this case the bones of large marine reptiles such as mosasaurs or plesiosaurs could have been used as a food source. The Oceanospirillales bacterial symbionts of Osedax produce collagenolytic enzymes and can live on cholesterol and collagen as primary carbon sources (8). Thus they should be able to feed on reptilian bones, although direct evidence of Osedax colonizing reptile bones is currently lacking. Furthermore, a Cretaceous origin seems possible because plesiosaur skeletons from late Cretaceous deep-water sediments were found associated with communities of small mollusks (22), showing that these skeletons could support invertebrate communities resembling those colonizing whale skeletons in the deep-sea today (23). Osedax has also been found living on experimentally submerged cow bones (24), indicating its ability to live on a variety of mammalian bones. To further test this hypothesis, Cretaceous marine reptiles should be investigated for similar traces (7).

Osedax has a known depth range of 30–3000 m (7) and can be found in dense colonies that are able to consume an entire whale skeleton within a few years (19). Molecular clock estimates and the early Oligocene boreholes presented here show that Osedax has been destroying bones for most of the evolutionary history of whales. Osedax has an impact on the taphonomy of whale skeletons that has significantly affected the quantity and especially the quality of fossil whales, as documented here. This previously unrecognized “Osedax effect” is especially well-illustrated in whale fossils that are preserved in deep-water strata where sedimentation rates are low, such as the bioeroded fossils from Washington State.

Materials and Methods

Materials.

The two fossil specimens were found in carbonate concretions derived from sediments of the Pysht Formation, which is exposed west of the mouth of Murdock Creek along the shore of the Strait of Juan de Fuca in Clallam County, Washington State. This is LACMIP locality 6295 and LACMVP locality 5412; also UWBM locality C0667. The age of the fossils from this locality (∼30 million years) is late early Oligocene on the basis of foraminiferans, mollusks, and mammal fossils (25) and magnetostratigraphy (26). Benthic Foraminifera, mollusks, large isopods, and the mode of preservation of whale fossils at this site, with the ventral side up and the limbs still in place, indicate that these sediments were deposited at bathyal depth (27, 28). In many cases, fossil bones from this locality were heavily corroded, probably bioeroded, before being fossilized (28), some to the point of being nothing more than amorphous fragments (29) with Osedax having likely played a major role. The bones were extracted from the enclosing concretion by etching with dilute (10% or less) formic acid. The figured specimen is deposited in the US National Museum of Natural History, Washington, DC (USNM 539939); the other specimen is deposited in the Burke Museum, Seattle, WA (UWBM 91837).

Micro-CT Scanning Method.

The x-ray microcomputed tomography scans of the whale bone that contains the traces of Osedax was done using the SkySkan1172 system (SkyScan) at the Institut für Geowissenschaften, Christian-Albrechts-Universität, Kiel, Germany. The fossilized bone was scanned with a beam energy of 100 kV, a flux of 100 μA, and a 0.5-mm-thick aluminum filter at a detector resolution of 8.7 μm/pixel using a 180° rotation with a step size of 0.6°. To study cavity morphology details, a second scan under identical beam conditions was performed, using a 360° rotation with a step size of 0.45° at a detector resolution of 8.0 μm.

Reconstruction, Segmentation, and Visualization.

The scan to survey the rib fragment (326 transmission images) was reconstructed in a 1,832 × 900 matrix of 2,029 slices with a resolution of 8.7 μm/voxel using the SkyScan software Nrecon running on a cluster of three networked PCs. The program uses a modified Feldkamp algorithm. The segmentation of bone and boreholes was done with the SkyScan software CT Analyzer. Before the subsequent generation of the 3D models by Adaptive Rendering, the data sets were downsized to one-third with an isotropic voxel resolution of 26.0 μm in a 610 × 300 matrix of 677 slices. Visualization of the 3D models was done by the SkyScan software CTvol. For the study of cavity morphology details, 800 transmission images were reconstructed in a 2,088 × 1,136 matrix of 1,998 slices. Segmentation and rendering were done on volumes of interest with a resolution of 8.0 μm/voxel.