146 JOURNAL OF M.H. EXPERIMENTAL SCHWEITZER ET ZOOLOGY AL. (MOL DEV EVOL) 285:146–157 (1999) Beta-Keratin Specific Immunological Reactivity in Feather-Like Structures of the Cretaceous Alvarezsaurid, Shuvuuia deserti M.H. SCHWEITZER,1,2* J.A. WATT,1 R. AVCI,3 L. KNAPP,4 L. CHIAPPE,5,6 M. NORELL,5,6 AND M. MARSHALL7 1 Department of Biology, Montana State University, Bozeman, Montana 59717 2 Museum of the Rockies, Montana State University, Bozeman, Montana 59717 3 ICAL Facility and Department of Physics, Montana State University, Bozeman, Montana 59717 4 Department of Biological Sciences, University of South Carolina, Columbia, South Carolina 29208 5 Department of Ornithology, American Museum of Natural History, New York, New York 10024 6 Department of Vertebrate Paleontology, American Museum of Natural History, New York, New York 10024 7 Walther Oncology Center, Indiana University, Indianapolis, Indiana 46202 ABSTRACT We report small fibrous structures associated with a new specimen of Shuvuuia deserti, which we hypothesize are remnants of feather-like epidermal appendages. Multiple analyses suggest that these structures are epidermally derived and contain epitopes consistent with βkeratin, a protein expressed only in extant “reptiles” and birds. Morphological, microscopic, mass spectrometric, and immunohistochemical studies are consistent with the interpretation that these structures are related to feathers. These data suggest that proteinaceous components may survive across geological time and support the view that alvarezsaurids (Shuvuuia and its allies) are either a lineage of birds or are a lineage phylogenetically close to them. J. Exp. Zool. (Mol. Dev. Evol.) 285:146–157, 1999. © 1999 Wiley-Liss, Inc. An exceptionally well-preserved skeleton of the alvarezsaurid Shuvuuia deserti (IGM 100/977) was recovered in 1993 in dune-derived fluvial sediments at Ukhaa Tolgod in southwestern Mongolia by the Mongolian-American Museum Expedition (Dashzeveg et al., ’95, Fig. 1a). Alvarezsaurids (Perle et al., ’93, ’94; Chiappe et al., ’96; Novas, ’96)1 are known from over 25 specimens at six localities throughout Late Cretaceous sediments of the Gobi Desert. Some of these specimens are complete and exhibit remarkable structural preservation, including well-preserved skulls and postcrania (Chiappe et al., ’98). Physical remnants of originally keratinous structures have been reported previously from other specimens (oviraptorosaurs) collected from Ukhaa Tolgod (Norell et al., ’95). During preparation of IGM 100/977, small white fibers of varying lengths and diameters were observed around the perimeter of the bones (Fig. 1b). These fibers were organized in small clumps and often lay nearly perpendicular to the axis of the bones. The fibers varied in length and width, averaging about 200 µm in diameter. The spatial orientation and arrangement of the extremely delicate fibers suggested that these might be remnants of feather-like structures. Furthermore, these fibers are hollow like the rami or barbs of all modern feathers. Figure 2 shows brightfield and darkfield micrographs of a 0.5 µm cross section of one of the fibers. With few and still controversial exceptions (Chen et al., ’98; Qiang et al., ’98), the appearance of feathers in the fossil record has been unambiguously linked to the presence of birds. 1 The type specimen of Mononykus olecranus is from Bugin Tsav, a putative Nemegt equivalent. Specimens collected at Ukhaa Tolgod can be diagnosed as separate, but closely allied taxon, Shuvuuia sp., recently described Parvicursor from Barun Goyot beds, consequently the collective mononykosaur is used here. IGM refers to Institute of Geology, Mongolia. Grant sponsor: National Science Foundation; Grant number: EAR 97-53187. *Correspondence to: M.H. Schweitzer, Dept. of Biology and Museum of the Rockies, Montana State University, Bozeman, MT 59717. E-mail: firstname.lastname@example.org Received 4 March 1999; Accepted 3 June 1999. © 1999 WILEY-LISS, INC. KERATIN IMMUNOREACTIVITY IN CRETACEOUS SPECIMEN 147 Fig. 1. (a) Photograph of Shuvuuia deserti during preparation showing articulated state and sediments surrounding the specimen. (b) Fibers surrounding the skeleton in situ. Recent feathers consist of keratin proteins, durable, hydrophobic filament proteins that, like other structural proteins (collagen, osteocalcin), are relatively resistant to degradation. Feathers are not susceptible to digestion by commonly occurring proteolytic enzymes such as trypsin, and must undergo some type of modification (e.g., disulfide bridge reduction, hydrolysis) for the degradation process to begin (Williams et al., ’90). For these reasons, keratin-derived structures, including feathers (Davis and Briggs, ’95; Norell et al., ’95; Chiappe et al., ’98; Schweitzer et al., ’99), are second only to bones and teeth in appearance in the fossil record (Logan et al., ’91). Extant feathers generally consist of a hollow calumus (root) and rachis (shaft), from which arise hollow barbs and from these, barbules and barbicels (Lucas and Stettenheim, ’72). Variants of the basic structure include down feathers, after- feathers and filoplumes, and the small, unbranched, bristle-like facial feathers that are seen near the eyes (Lucas and Stettenheim, ’72). The shape and distribution of these variants are related to their role in aerodynamic function (Parakkal and Alexander, ’72), or insulating and/or waterproofing capacity. The complexity of feather structure is reduced by varying degrees in most birds that have become secondarily flightless, and is least constrained in ratites, which have been flightless the longest (McGowan, ’89). On the molecular level, mature feathers are the result of the controlled expression of a family of genes that code for the keratin proteins. Although there are no “feather” genes, the diverse morphologies of feathers in a single specimen or among different bird taxa are the result of variation in timing and expression of a small group of homologous and relatively homogeneous beta keratin 148 M.H. SCHWEITZER ET AL. Fig. 2. Dark-field (a) and bright-field (b) micrographs of fiber in cross section. Arrow indicates hollow central region. For comparison, a cross section of a duck feather barbule is seen in fluorescent image in Figure 4. Bar = 40 µm. genes (Gregg et al., ’83; Gregg and Rogers, ’86; Brush, ’93, ’96). Although alpha keratin is expressed in the epithelial tissues of all vertebrates (Sawyer et al., ’86; Shames and Sawyer, ’87), the beta keratin structural proteins are unique to nonavian reptiles (crocodiles, turtles and lepidosaurs) and birds among extant vertebrates. Alpha keratins can be distinguished from the more restricted beta keratins by a greater molecular weight (40,000– 70,000), and protein filament diameters of 8–14 nm. Beta keratins can be divided into two groups. The larger (MW 14,000–25,000) are isolated from scale, claw, and beak tissues. Feather proteins are smaller (MW ~10,000), as a result of a 50 bp deletion in the original beta keratin gene (Gregg and Rogers, ’86), although the filament diameter in both groups is 3–4 nm. In feathers, beta keratin filaments occur in bundles oriented along the long axis of the feather, a different pattern than that observed in other epidermal structures in both nonavian reptiles and birds (Parakkal and Alexander, ’72). Thus, while both alpha and beta keratins are essential components of the differentiation of epidermal structures of nonavian reptiles and birds, mature feathers are unique among epidermal structures in that they consist only of beta keratin. to avoid contamination. Treatment of modern controls was identical to that of fossil tissues for all analyses. All experiments were repeated multiple times to validate results. The fibers associated with this specimen of Shuvuuia were removed from the sediment matrix by saturating both with 100% ethanol. Once isolated, the fibers were allowed to infiltrate overnight or for several days at 4°C in LR White (hard grade) embedding medium (Ted Pella) designed for immunohistochemical studies. The infiltrate was removed, fresh medium applied, and allowed to polymerize overnight at 60°C. Sections were cut with dedicated glass knives to a thickness of 0.5– 1.0 µm using a Sorvall Ultramicrotome, and mounted to glass slides coated with a gelatin adhesive. Sectioning was difficult in some cases because the sample tended to pull away from the embedding material. Samples of duck feathers were infiltrated and polymerized in tandem with the fibers and sectioned with glass knives kept separate from those used on the ancient material. Tissues were not fixed in either case because it was felt that chemical treatment of the ancient material should be kept to a minimum to avoid alteration of any protein epitopes possibly preserved. EXPERIMENTAL PROCEDURES Tissue handling and preparation At no time were any of the samples handled directly, and instruments used in analyses were either new and dedicated or sterilized before use Scanning electron microscopy Scanning electron microscopy (SEM) was performed on the ancient fibers associated with IGM 100/977 and compared with feathers from extant birds. Unembedded sections of all specimens were placed on double-sided, carbon tape as whole KERATIN IMMUNOREACTIVITY IN CRETACEOUS SPECIMEN mounts, coated with gold-palladium (10 Å) and visualized using a JEOL 6100 scanning EM at 10–15 kV. Transmission electron microscropy Ancient and modern samples were infiltrated and polymerized in LR White, as described. Ultrathin sections of approximately 90–95 nm were collected on formvar-coated slot grids, counterstained with uranyl acetate and lead citrate, then viewed using a Zeiss EM10C transmission electron microscope at 60–80 kV. 149 the amount of specific antibodies in the serum was used to control for spurious binding due to concentration effects. Other controls are outlined in Schweitzer et al. (’99). Coverslips were applied to the sections using Vectashield mounting medium (Vector) and viewed using a BioRad DVC 250 confocal microscope equipped with an argonkrypton laser and a Photonics cooled color integrating CCD camera. All data were integrated over 1/8 second, and images were captured using NIH Image software. Immunochemistry Time of flight secondary ion mass spectroscopy (TOF-SIMS) Antisera used in this study were developed by O’Guin et al. (’82) to distinguish between alpha and beta keratins. The specificity of these antisera have been demonstrated by double diffusion assays, indirect immunofluorescence, and immunoblot analyses (O’Guin et al., ’82; Knapp et al., ’91). This specificity is further supported by our own experiments preabsorbing the sera against minced feathers. Following this incubation, preabsorbed antisera were incubated with the test samples, and resulting immunofluorescence was almost eliminated. All incubation steps were performed at 4°C and separated by at least three 15-min PBS washes. All sections were initially treated with 1 mg/ml sodium borohydride (NaBH4) to expose epitopes and reduce autofluorescence. Specimens were then incubated for 1 hr in 4% normal goat serum to reduce nonspecific staining, then incubated with either rabbit anti-avian alpha keratin (1:100) or rabbit anti-avian beta keratin (1:50) at 4°C overnight. Sections were then incubated sequentially with biotinylated goat anti-rabbit IgG (1:200 for 1 hr, Vector) and conjugated Avidin-CY5 (1:1000, Jackson ImmunoResearch) or Avidin-DCS conjugated to fluorescein isothiocyanate (FITC) (1:1000, Vector) for 1.5–2.0 hr. Sections of duck feather were processed in an identical manner to the positive controls. Negative controls of both duck feather and ancient fiber were identical to the above, with substitution of the anti-keratin primary with the nonrelevant rabbit anti-human ACTH (adrenocorticotropin hormone, 1:50, to control for nonspecific Ig binding) or normal rabbit sera (1:50, to control for nonspecific reaction with serum components unrelated to the presence of primary antibodies) in place of the specific primary antisera. Incubation with sera preabsorbed against feathers to reduce A Charles Evans and Associates (now PHIEvans) TRIFT 1 system was used to collect the TOF-SIMS spectra. Primary ions were produced by a gallium gun with an average energy of 15 keV. A gallium pulse of width ~1 ns bombarded the sample with a repetition rate of 10 kHz. The ion bean was rastered over the 80 × 80 µm area, and total acquisition time for each sample was 5 min. The gallium primary ions generate low energy ions (<1 eV) of molecular fragments (secondary ions), which are accelerated to ~3000 eV. Only a small portion of the desorbed fragments are charged, and these are the source of mass data. Masses of the fragments are determined by accurate measurements of the time of flight of these ions from the sample to the detector. Small fragments (~200–400 µm diameter) of IGM 100/977 fibers were pressed against clean gold-coated, indium foil using a specially cleaned glass slide. A fragment of feather of approximately the same size as the ancient sample was treated in an identical manner and used as a positive control. Grains of sand from the sediments surrounding the fiber, a section of duck feather rachis and solubilized human keratin (sigma) were analyzed as controls. However mass resolution, defined as mass divided by mass uncertainty, is highly matrix dependent; therefore patterns and distributions of peaks vary greatly across samples that exhibit different matrices. The assignments of protein fragments to TOFSIMS masses have been well described in the literature (e.g., Mantus et al., ’93; Lhoest and Gastner, ’98; Lhoest and Gastner, pers. comm.), and the high mass resolution identifies each peak uniquely and with certainty. The areas under the peaks corresponding to amino acid fragments are normalized to the integrated C2H3+ peak area (265 counts/sec) (Galuska, ’94; pers. comm.). Antisera 150 M.H. SCHWEITZER ET AL. RESULTS The morphology, orientation, and association of the fibers with the skeletal remains of IGM 100/ 977 suggested an organic rather than geological source. Two obvious sources of contamination that might give rise to structures of similar gross appearance include plant root structures or fungal hyphae. However, gross microscopic study of the fibers was not consistent with either plant material (no cell walls visualized) or fungi (no segmentation, branching, or evidence of spores or fruiting bodies). Fibers were embedded, sectioned, and subjected to digestion with cellulase, an enzyme that degrades plant material by specifically digesting the cellulose in plant cell walls. After a 24-hr incubation with the enzyme, no alteration of the microstructure of the fibers was visualized, although the cell walls of extant plant material treated in parallel were degraded. Additionally, 1-µm sections of the fibers subjected to staining with cotton blue, a common screening agent for fungi, were negative. Scanning electron micrographs of the fibers were then compared with similar images of barbs from duck contour feathers and specialized head feathers of Cathartes aura, the turkey vulture. Figure 3A and B shows a comparison of one of the fibers associated with Shuvuuia and that of the vulture. The shafts of both specimens are rounded, lacking the longitudinal furrow typical of flight feathers. Additionally, there are no obvious barbs or barbules visualized in either sample. The vulture feather possesses a series of platelike scales near the base (point of insertion into the skin), and a slight flattening of the feather shape is visualized. This pattern is similar to that seen in the Shuvuuia fibers. Figure 3C, at higher magnifications, shows a smooth outer surface on the vulture feather, which covers a more rugose inner structure, possibly pith. Again, a similar pattern can be seen in the fiber from Shuvuuia (Fig. 3D). Finally, very small hair-like projections were visualized on the fiber surface (Fig. 3F, arrow) and compared with like structures projecting from the barbule of a duck flight feather (Fig. 3E). The similarity of these two projecting structures at high magnifications is notable. Since keratin microfilaments are of a known and consistent diameter, ultrathin sections (~90–95 nm) of the ancient sample and duck feather were visualized using transmission electron microscopy. The duck feather shows the microfilaments, either singly or in bundles, embedded in what appears to be an amorphous protein matrix (Fig. 4A). In the ancient sample, microfilaments are seen localized to “pockets” of an electron dense mineral matrix (Fig. 4B). Like the duck feather, however, the microfilaments in the ancient sample are either seen singly or in bundles, and are of a diameter consistent with beta keratin (Parakkal and Alexander, ’72). Structural observations were supplemented with immunohistochemical studies. Antibody recognition of small, three-dimensional regions of proteins, called epitopes (3–5 amino acid residues; Child and Pollard, ’92), is highly specific and as such is routinely used in medical diagnostics or embryological or ontogenetic studies to localize specific protein expression in tissues. Because of its specificity, antibody-antigen recognition does not depend on the presence of total protein, and can be used to identify protein fragments (Child and Pollard, ’92). It is much more plausible that very small fragments of proteins may be preserved across geological time than total proteins. This, coupled with the ability to localize protein signal to specific tissues, makes immunohistochemical techniques promising for use with ancient specimens. When sections of Shuvuuia fibers were incubated with primary antibodies raised against extant avian alpha and beta keratins (O’Guin et al., ’82), and compared with sections of duck feather, treated in tandem with the ancient specimens, both the modern and ancient samples show very slight reactivity with the alpha keratin antibodies (Fig. 5a, b), which may be the result of “edge effect,” or nonspecific binding of antibodies to roughened edges of sectioned material. However, tissue reaction to beta keratin antibodies is significantly greater in both samples, indicating strong reactivity of both the ancient and modern samples to these specific antibodies (Fig. 5c, d). When sections were incubated with either normal (nonimmunized) sera (containing antibodies, but none specific for beta keratin) or with an antiserum against a non-relevant protein such as human anti-ACTH (adrenocorticotropin hormone) at the same dilutions as the test sera, reaction above background was not noted in either specimen (Fig. 5e, f). When beta keratin antisera were first incubated with minced feathers to reduce the concentration of specific antibodies, and then exposed to fiber sections as described, the fluorescent staining was reduced to background, indicating that the reaction to the beta keratin antibodies is specific and not artifact (Fig. 5g). Finally, to verify that amino acid residues consistent with proteinaceous components were KERATIN IMMUNOREACTIVITY IN CRETACEOUS SPECIMEN 151 Fig. 3. Scanning electron micrographs of Shuvuuia fiber in comparison with extant samples. (A) Low magnification showing the cylindrical shape and lack of branching beyond the base of turkey vulture bristle. (B) IGM 100/977 at the same magnification shows a similar overall morphology. Note also in both specimens the plate-like structures near one end. Higher magnification of vulture feather (C) and IGM 100/ 977 (D) shows both structures to have a smooth outer covering, and exposes a rugose, or fibrous interior, consistent with pith. (E) Small, hair-like projection on the surface of a rachis from a duck feather (F). Similar projection is seen on the Shuvuuia fiber. This projection from the duck feather (G) and IGM 100/977 (H) is magnified, once again demonstrating overall similarity of structure. present in these fibers, unembedded samples of the fiber were subjected to time of flight secondary ion mass spectroscopy (TOF-SIMS). This method was chosen for a number of reasons, including sensitivity (detection to submonolayers), minimal requirement for sample size, and the ability to resolve molecules that differ by as little as 5–10 milliatomic mass units (amu). Most important was the ability to apply this method to wholesample analyses, which eliminated the need to subject the very small fibers to chemical extraction techniques which could introduce contamination or cause loss of material through the various steps involved in extraction and purification. Table 152 M.H. SCHWEITZER ET AL. Fig. 4. Transmission electron micrographs of 95-nm sections of duck feather (A) and Shuvuuia fiber (B). In (A) thin filaments or filament bundles are embedded in a protein matrix that appears amorphous at this magnification, but which is actually an extension of the keratin molecules. Note a single protein filament of ~3 nm in diameter. In (B), filaments or bundles of filaments of similar diameter are seen more randomly oriented. Single filaments measure 3–4 nm in diameter. Lower magnifications (not shown) reveal these filaments to be encased in small “pockets” of electron dense mineral matrix. Original scope magnification: A, 40,000; B, 50,000. 1 shows the amino acid fragments identified in a fragment of the Shuvuuia fiber by this technique. The numbers in parentheses next to the chemical formulae show the percent intensities relative to C2H3+ peak, which allows a relative quantification of the organic fragments. Figure 6 shows a portion of the positive ion spectra from which the table is constructed. Data are consistent with the interpretation that amino acids are present in this fiber sample. However, because this sample is not purified, is highly degraded, and contains inorganic components (minerals), it is not possible to determine amino acid ratios. Additionally, TOFSIMS is highly matrix dependent, and unless two samples have similar matrices, it is not possible to compare them. epidermally derived structures which are similar to feathers on a molecular level. These fibers are seen to be hollow, both grossly and in microscopic cross section, which is consistent with feather structure. However, they are morphologically dissimilar to modern flight or contour feathers in that they show no evidence of the complex branching patterns characteristic of all modern bird body feathers. This may be a diagenetic effect, thus structural similarity to the highly specialized and derived head feathers of the vulture may be only superficial. Alternatively, this may reflect the original structural pattern of these fibers. If reflective of original structure, these fibers suggest an intriguing insight into the evolution of feathers, and may support the idea that the morphological evolution of feathers followed that seen in ontogenetic patterns, where the earliest feather structures to arise from placodes resemble small simple hollow filaments (Brush, ’99). Additionally, this would provide support for the hypothesis that feathers as an evolutionary novelty did not arise for flight. Although the data from immunological studies suggest that the antiserum raised against avian beta keratins specifically recognizes components within the ancient fiber, the fibers have been in- DISCUSSION Multiple lines of evidence suggest that the small fibers found in association with IGM 100/977 contain organic components that are consistent with degradation products of the protein beta keratin. Since this protein is only produced by epidermal cells of nonavian reptiles and birds, and since feathers are the only epidermal structures in modern taxa that consist of only beta keratin, it is hypothesized that these fibers may represent KERATIN IMMUNOREACTIVITY IN CRETACEOUS SPECIMEN 153 Fig. 5. Immunohistochemical results. (a) Shuvuuia and (b) duck feather exposed to avian anti-alpha keratin primary antisera at a 1:100 dilution. (c) Shuvuuia and (d) duck feather exposed to avian anti-beta keratin primary antisera, at a dilution of 1:50. (e) Shuvuuia and (f) duck feather exposed to normal (nonimmunized) rabbit sera at 1:50 dilution. (g) Shuvuuia exposed to avian anti-beta keratin antisera (1:50 dilution) which had been previously incubated repeatedly with minced feather to significantly reduce the concentration of antibodies specific to beta keratin. The signal is reduced to background, demonstrating that the antibodies binding to the fiber are blocked by exposure to beta keratin, thus supporting the specificity of these antibodies for beta keratin proteins. For all panels, intensity of fluorescent signal is directly correlated with the degree of antibody binding to epitopes in the exposed tissues. For all images, bar = 20 µm. filtrated with minerals. Therefore, one must entertain the possibility that this recognition arises from nonspecific interaction between the antisera and the mineral. If this were the case, the control sera should demonstrate equal reactivity, which is not the case. Another alternative is that characteristics of the protein epitopes were preserved intact by the fossilization process in a manner similar to molecular imprinting (Mosbach and Ramstrom, ’96; Huaiquiu et al., ’99), without preserving the molecules themselves—although it is difficult to understand how this could occur in mineral with sufficient accuracy for antigenic recognition to be maintained. Finally, although these mass spectrometry data are not sufficient to conclude that identified amino acids are endogenous 154 M.H. SCHWEITZER ET AL. TABLE 1. Amino acid fragments identified in a fragment of the Shuvuuia fiber AA ser gly pro ala cys val leu glu asp thr phe lys met gly-gly ala-gly ala-ala 1 2 Positive ion fragments (% intensity relative to C2H3) CH3O (18.7) CH4N (146.2) C4H8N (3.1) C2H6N (25.8) CH3S (4.4) C4H10N (23.1) C5H12N (35.9) C2H3O (50.0) C2H6N (23.9) C2H5O (14.6)1 C6H5 (8.8)2 C4H11N (8.3) C3H7S (14.1) C4H7N2O2 (3.0) C5H9N2O2 (1.1) C6H11N2O2 (0.7) C2H6NO (0.9) C2H4N (14.9) C4H6N (2.7) CH2N (12.2) C4H7N2O2 (3.0) C3H6N (18.5) C3H8N (15.7) C3H8NO (0.4) C7H7 (6.1)** C5H10N (3.9) C2H5S (1.7) C3H5O (9.5) C8H10N (0.9) The R group of thr C2H5O+ (CH3=+CH-OH) cannot be differentiated from the environmental contaminants +CH2-CH2-OH or CH3-O-+CH2. Hydrocarbon fragments are common to most surfaces and can be due to environmental contaminants. to the specimen, their identification in and localization to whole sample analyses of the ancient fibers supports the possibility that these fibers are remnants of the original epidermal structures. In addition, clusters of nitrogen-containing saturated hydrocarbons were identified in IGM 100/ 977 (Fig. 6b). Peaks shown in the upper panel are uniquely identified; however the peaks of greater mass in the lower panel are labeled as to best fit. These molecules may arise from protein degradation and subsequent condensation reactions (Collins et al., ’98). As mentioned above, TOF-SIMS does not allow us to state conclusively that all amino acid fragments are endogenous to the sample, but the presence of amino acid fragments, peptide linked fragments, and some high molecular weight molecules does support the hypothesis that some original molecular components remain in these fibers. This study supports several conclusions: (1) It is the first report of feather-like structures preserved in three dimensions, not as impressions, from an eolian depositional environment. (2) The identification of epidermally derived structures that are molecularly consistent with feathers in so ancient a lineage may shed light on the evolution of feather structure, flight and endothermy in Cretaceous organisms. (3) It supports the hypothesis that informative protein components may be identified across geological time spans. (4) It advances the possibility of using immunohistochemical techniques to identify protein fragments and organic components in other fossil specimens, such as the exceptionally well-pre- served specimens of Sinosauropteryx (Chen et al., ’98), Caudipteryx (Qiang et al., ’98) and Protoarchaeopteryx (Qiang et al., ’98), and demonstrates the utility of this method in cases where the number of samples is extremely limited. (5) It establishes the use of TOF-SIMS as an effective method for molecular mass analysis of organic components of fossils without destruction or in cases where very little material is available. Finally, this research suggests further study into the nature of the antigens to which the antisera bind. If the antigen is not organic but a geologically derived molecular mimic, it may be possible to obtain usable molecular information even though the original organic structures are no longer present. The occurrence of feathers in association with alvarezsarid specimens is congruent with the hypothesis of avian relationships of this lineage (Perle et al., ’93; Chiappe et al., ’96; Novas, ’96; Forster et al., ’98). Nevertheless, since feathers have been documented in nonavian maniraptoriforms (Qiang et al., ’98), their presence in Shuvuuia does not provide any additional evidence in support of such a phylogenetic hypothesis. The phylogenetic placement of the Alvarezsauridae is controversial. Alvarezsaurids are highly specialized creatures and the rapid rate of new discoveries of early birds and non-avian maniraptoriforms has kept phylogenetic hypotheses about this segment of theropod evolution highly dynamic (Fig. 7). If alvarezsaurids are regarded as nonavian theropods, the identification of the structures surrounding the skeleton of Shuvuuia adds additional evidence to the idea that feathers were common to most, or all, nonavian maniraptoriforms. KERATIN IMMUNOREACTIVITY IN CRETACEOUS SPECIMEN 155 Fig. 6. (a) Positive ion spectra of Shuvuuia in the region from 0 to 200 amu (top panel). Note the inorganic components that dominate the spectra in the low mass regions (0–50 amu). Additionally, nitrogen-containing hydrocarbons (e.g., 130 amu) are identified in the fiber. Middle panel is expansion of the spectra in the region of 60–130 amu. Peaks representing some of the amino acid fragments in this sample are as labeled. Bottom panel is a further expansion of the region 71.85–72.20 amu. Resolution of this peak is unambiguous. (b) Positive ion spectra in the region from 165 to 300 amu (top panel) identifies additional organic components in the fiber in the higher mass regions, showing clusters of nitrogen-containing saturated hydrocarbons. We hypothesize that these are derived from nitrogen-containing parent molecules because they are separated by CH2 units and contain a single amino group. Bottom panel shows the region between 350 and 500 amu. These large peaks may correspond to parent molecules alluded to above. Molecules containing one nitrogen are even in mass, while odd masses contain two nitrogen groups (e.g., mass 183 with formula C11H23N2, mass 200 is C14H18N). 156 M.H. SCHWEITZER ET AL. with feathers can be identified in nonavian theropods, the phylogenetic proximity of these lineages to birds is robustly supported. (2) The morphological evolution of feathers may be further elucidated if it can be shown that the first epidermal structures to appear were simple structures that could not have functioned in flight and that flight feathers resulted from selection for increasingly complex forms. If beta-keratin specific epidermal structures that were unable to be classified morphologically as feathers are identified, it would imply that the molecular mechanism and expression of feather proteins existed before feathers arose as morphologically distinct structures in their modern form. (3) The only organisms today that possess epidermally derived body coverings are those with basal metabolic rates significantly above ectothermy. Demonstration of such a body covering in nonavian theropods implies that they likewise possessed a metabolic strategy above ectothermy. Further application of immunological techniques to suitable fossil specimens, such as use of monoclonal antibodies, phage display libraries, or epitope mapping techniques may be employed to aid in the elucidation of phylogenetic hypotheses (Collins et al., ’91; Lowenstein and Scheuenstuhl, ’91; Lowenstein, ’93). ACKNOWLEDGMENTS This research was funded by donations from Dr. G. Ellis. We thank Dr. C. Paden, in whose lab the immunohistochemical studies were done. In addition, we thank N. Equall, D. Gustafsen, C. Horner, and A. Hageston for technical expertise and J. Clark and A. Davidson for their help identifying samples. Finally, our gratitude goes to S. Pincus, C. Forster, J. Starkey, M. Tientze, A. Brush, B.A. Stankiewicz, R. Sawyer, K. Padian and M. Collins for their patience, critiques, reviews, ideas, support and helpful discussions. Fig. 7. Possible alternatives for phylogenetic placement of the Alverezsauridae (Shuvuuia and its relatives). (A) Alvarezsaurids are considered to be birds, as the sister-taxon to all birds other than Archaeopteryx (Perle et al., ’93; Chiappe et al., ’96; Novas, ’96). (B) Alvarezsaurids are regarded to be the sister-taxon to birds (Chiappe, ’99). (C) Alvarezsaurids are placed in a more basal position within coelurosaurian theropods (Novas, ’96; Sereno, ’97). The presence of feathers in nonavian theropods is significant from evolutionary and physiological perspectives for several reasons: (1) Feathers have been used to uniquely identify birds; however, if integumentary structures molecularly consistent LITERATURE CITED Brush AH. 1993. The origin of feathers: a novel approach. Avian Biol 9:121–161. Brush AH. 1996. On the origin of feathers. J Evol Biol 9:131–142. Brush A. 1999. In: Wolberg D, editor. Dinofest III: proceedings (in press). Chen P-J, Dong Z, Zhen S-N. 1998. An exceptionally wellpreserved theropod dinosaur from the Yixian Formation of China. Nature 391:147–152. Chiappe L, Norell MA, Clark JM. 1996. Phylogenetic position of Mononykus (Aves: Alvarezsauridae) from the Late Cretaceous of the Gobi Desert. Mem Queens Mus 39:557–582. Chiappe LM. 1999. The phylogeny of basal birds. In: Ostrom symposium volume (in press). KERATIN IMMUNOREACTIVITY IN CRETACEOUS SPECIMEN Chiappe LM, Norell MA, Clark JM. 1998. The skull of a relative of the stem-group bird Mononykus. Nature 392: 275–278. Child A, Pollard M. 1992. A review of the applications of immunochemistry to archaeological bone. J Arch Sci 19:39–47. Collins MJ, Muyzer G, Westbroek P, Curry GB, Sandberg PA, Xu SJ, Quinn R, MacKinnon D. 1991. Preservation of fossil biopolymeric structures: conclusive immunological evidence. Geochim Cosmochim Acta 55:2253–2257. Collins MJ, Walton D, King A. 1998. The geochemical fate of proteins. In: Stankiewicz BA, van Bergen P, editors. Nitrogen-containing macromolecules in the bio- and geosphere. Am Chem Soc. 370 p. Dashzeveg D, Novacek MJ, Norell MA, Clark JM, Chiappe LM, Davidson A, McKenna MC, Dingus L, Swisher C, Perle A. 1995. Unusual preservation in a new vertebrate assemblage from the Late Cretaceous of Mongolia. Nature 374:446–449. Davis PG, Briggs DEG. 1995. Fossilization of feathers. Geology 23:783–786. Forster CA, Sampson SD, Chiappe LM, Krause DW. 1998. The theropodan ancestry of birds: new evidence from the Late Cretaceous of Madagascar. Science 279:1915–1919. Galuska A. 1994. Surface characterization of EVA co-polymers and blends using XPS and TOF-SIMS. Surf Interface Anal 21:703. Gregg K, Rogers GE. 1986. Feather keratin: composition, structure and biogenesis. In: Bereiter-Hahn J, Matoltsy AG, Richards KS, editors. Biology of the integument, vol. 2: vertebrates. Berlin: Springer-Verlag. p 667–694. Gregg K, Wilton SD, Rogers GE, Molloy PL. 1983. Avian keratin genes: organization and evolutionary inter-relationships. In: Nagley P, Linnane AW, Peacock WJ, Pateman JA, editors. Manipulation and expression of genes in eukaryotes. New York: Academic Press. p 65–72. Huaiquiu Shi, Tsai, Wei-Bor, Garrison M, Ferrari S, Ratner BD. 1999. Template-imprinted nanostructured surfaces for protein recognition. Nature 398:593–597. Knapp LW, Linser PJ, Carver WE, Sawyer RH. 1991. Biochemical identification and immunological localization of two non-keratin polypeptides associated with the terminal differentiation of avian scale epidermis. Cell Tiss Res 265:535–545. Lhoest J-B, Gastner DG. 1998. Time-of-flight SIMS studies of adsorbed protein mixtures. In: Proceedings of the 11th annual SIMS workshop. Austin, TX. p 41. Logan GA, Collins MJ, Eglinton G. 1991. Preservation of biomolecules. In: Allison PA, Briggs DEG, editors. Taphonomy: releasing the data locked in the fossil record. New York: Plenum Press. p 1–24. Lowenstein JM. 1993. Immunospecificity of fossil proteins. In: Engel MH, Macko SA, editors. Organic Geochemisty. New York: Plenum Press. p 817–827. Lowenstein JM, Scheuenstuhl G. 1991. Immunological meth- 157 ods in molecular palaeontology. Phil Trans R Soc Lond B 333:375–380. Lucas AM, Stettenheim PR. 1972. Avian anatomy: the integument, part I: Agricultural Handbook 362. Washington DC: U.S. Dept. of Agriculture. 340 p. Mantus DS, Ratner BD, Carlson BA, Moulder JF. 1993. Static secondary ion mass spectrometry of adsorbed proteins. Anal Chem 65:1431–1438. McGowan C. 1989. Feather structure in flightless birds and its bearing on the question of the origin of feathers. J Zool Lond 218:537–547. Mosbach K, Ramstrom O. 1996. The emerging technique of molecular imprinting, and its future impact on biotechnology. Biotech 14:163–170. Norell MA, Clark JM, Chiappe LM, Dashzeveg D. 1995. A nesting dinosaur. Nature 378:774–776. Novas FE. 1996. Alvarezsauridae: Cretaceous maniraptorans from Patagonia and Mongolia. Mem Queens Mus 39:675–702. O’Guin WM, Knapp LW, Sawyer R.H. 1982. Biochemical and immunohistochemical localization of alpha and beta keratin in avian scutate scales. J Exp Zool 220:371–376. Parakkal PF, Alexander NJ. 1972. Avian epidermis. In: Keratinization: a survey of vertebrate epithelia. New York: Academic Press. p 26–57. Perle A, Norell MA, Chiappe LM, Clark JM. 1993. Flightless bird from the Cretaceous of Mongolia. Nature 362: 623–626. Perle A, Chiappe LM, Clark JM, Barsbold R, Norell MA. 1994. Skeletal morphology of Mononykus olecranus (Theropoda: Avialae) from the Late Cretaceous of Mongolia. Novitiates 3105:1–29. Qiang J, Currie PJ, Norell MA, Shu-An J. 1998. Two feathered dinosaurs from northeastern China Nature 393:753–761. Sawyer RH, Knapp LW, O’Guin WM. 1986. The skin of birds: epidermis,dermis and appendages. In: BereiterHahn J, Matoltsy AG, Richards KS, editors. Biology of the integument, vol 2: vertebrates. Berlin: SpringerVerlag. p 194–234. Schroeder RA, Bada JL. 1976. A review of the geochemical applications of the amino acid racemization reaction. EarthSci Rev 12:347–391. Schweitzer MH, Watt JA, Avci R, Forster C, Krause DW, Knapp L, Rogers RR, Beech I, Marshall M. 1999. Keratin-specific immunoreactivity in claw sheath material from a late Cretaceous bird from Madagascar. J Vert Paleo (in press). Sereno PC. 1997. The origin and evolution of dinosaurs. Ann Rev Earth Planet Sci 25:435–490. Shames RB, Sawyer RH. 1987. Expression of b-keratin genes during development of avian skin appendages. Curr Top Dev Biol 22:235–253. Williams CM, Richter CS, MacKenzie JM Jr, Shih JCH. 1990. Isolation, identification and characterization of a featherdegrading bacterium. Appl Env Microsc 56:1509–1515.