Matrix from MOR 91.72 was more consistent in morphology and robustness with past reports on extant collagenous matrix (Schweitzer et al., 20052013) than what was observed for matrix of MOR 604 and MOR 605. “Collagenous” matrix of MOR 604 and MOR 605 was closer in morphology to what has previously been reported for Mesozoic dinosaurs (Schweitzer et al., 20052009; Schweitzer, Wittmeyer, & Horner, 2007) and early–mid Cenozoic organisms (Cadena, 20162020). The stark difference in these observations supports a disparity in degree of type-1 collagen preservation between these specimens, which is predicted to affect potential sequencing analyses. Prior studies have reported type-1 collagen sequences from MOR 604 (Schweitzer et al., 2002) and MOR 605 (Asara et al., 2007). MOR 91.72 however has not previously been sequenced and a direct comparison regarding degree of type-1 collagen sequence preservation is not currently possible. Further, MOR 91.72, MOR 604, and MOR 605 were all recovered from the same geographic region, albeit different burial sites (Schweitzer, Wittmeyer, & Horner, 2007). This supports the observed dichotomy in “collagenous” matrix preservation is thus likely less dependent on thermal setting.
Another study that analyzed biomolecular histology to a limited extent is that of the Pliocene Ellesmere Island camel tibia (Rybczynski et al., 2013). A cross-section of a vascular canal within the tibia was elementally mapped using energy dispersive X-ray spectroscopy (EDS). The analysis demonstrated that elements consistent with iron oxyhydroxides and barium sulfates colocalized to the vascular canal. The presence of such exogenous minerals supports that this tibia had undergone substantial chemical alteration. Both mineral precipitants are consistent with observations from older tertiary (Boskovic et al., 2021; Cadena, 20162020) and even Mesozoic specimens (Armitage & Anderson, 2013; Boatman et al., 2019; Schweitzer et al., 20132014; Surmik et al., 2016), and their presence likely precludes it from being considered a “subfossil.” Despite the apparent chemical alteration to its biomolecular histology, the tibia still preserved collagen sequences identifiable via mass spectrometry (Rybczynski et al., 2013).
Samples from the Pliocene tibia were not demineralized and examined with light microscopy within the study (Rybczynski et al., 2013), however, thus precluding a direct comparison against observations for the MOR 91.72, MOR 604, and MOR 605 “collagenous” matrix morphology. The substantial mineralization detected by the EDS analysis is consistent with observations of mineralized histological structures within MOR 604 and MOR 605 (Schweitzer, Wittmeyer, & Horner, 2007). This supports a hypothesis that any “collagenous” matrix the Ellesmere Island tibia preserves is likely highly degraded morphologically, in a manner consistent with MOR 604 and MOR 605 (Schweitzer, Wittmeyer, & Horner, 2007) as well as previous reports for Mesozoic dinosaurs (Schweitzer, Wittmeyer, & Horner, 2007) and pre-Pliocene Cenozoic (Boskovic et al., 2021; Cadena, 20162020) specimens.
Data from the two studies above already enables some predictions to be made regarding the relationship of the specimens' underlying biomolecular histology with degree of sequence preservation, and even some diagenetic variables. The dichotomy in biomolecular histology between extant specimens along with MOR 91.72 when compared against MOR 604, MOR 605, the Pliocene camel tibia, and Mesozoic dinosaurs is to this point a largely unexplored finding. Few, if any, studies have directly explored how these differences in demineralized “collagenous” matrix histology manifest in degree of recovered sequence data.
Based on the discussion above, bone specimens preserving ancient DNA are herein hypothesized to possess an intact, robust collagenous matrix somewhat similar to that of MOR 91.72. If sequence-able DNA is present, collagenous matrix would also still be expected to be relatively intact. In contrast, bone specimens with a brittle, easily fragmented “collagenous” matrix like that of MOR 604 and MOR 605 are predicted to preserve, at most, remnant peptide sequences. If the collagenous matrix has degraded to the point it has lost much of its structural integrity, the preservation of sequenceable DNA is not expected (Briggs et al., 2000; Lindahl, 1993; Wang et al., 2012). Further, this agrees with the trend of sequenceable DNA being rarely reported from specimens exceeding 0.13–0.24 Ma in geologic age (excluding cave and permafrost deposits) (Buckley et al., 2011; Froese et al., 2017; Lindqvist et al., 2010; Meyer et al., 2017; Mitchell & Rawlence, 2021; Wadsworth & Buckley, 2014; Welker et al., 2019) as MOR 604 and MOR 605 are both assigned constrained dates of ~100–600 Ka (Hill & Schweitzer, 1999; McDonald et al., 2020; Wilson & Hill, 2000).
If such a hypothesis were supported, practical methods such as electron and even light microscopy may be capable of screening fossil/subfossil specimens for sequence preservation with high precision. However, the limited extent of the data that has been reported for ancient vertebrate biomolecular histology severely limits the conclusions that can be drawn regarding these relationships. This epitomizes the need emphasized by this review for extensive study of fossil/sub-fossil vertebrate biomolecular histology.

5 METHODS FOR STUDYING FOSSIL/SUBFOSSIL BIOMOLECULAR HISTOLOGY

Examination of fossil/subfossil biomolecular histology is proposed for empirically studying how the cumulative effect of diagenetic variables upon a specimen's biomolecular histology correlates with degree of sequence preservation. Vertebrate elements with the highest potential for molecular sequence preservation include tooth enamel and dentine, bone, and eggshell (Demarchi et al., 2016; Wang et al., 2012; Welker et al., 2019). Of these, bone by far is the most widely characterized within ancient specimens as to its nonmineral histological structures. Numerous studies have reported histological structures morphologically and chemically consistent with biological cells, vascular tissue, and “collagenous” matrix preserved within Cenozoic and Mesozoic bones (Armitage & Anderson, 2013; Boatman et al., 2019; Boskovic et al., 2021; Cadena, 20162020; Lindgren et al., 2011201520172018; Schweitzer et al., 20052009; Schweitzer, Wittmeyer, & Horner, 2007; Surmik et al., 2016; Wiemann et al., 2018). In particular, the organic portion of extant collagenous bone matrix is comprised of ~90% type-1 bone collagen (Boatman et al., 2019; Wang et al., 2012). This high proportion of a single, specific molecule is practical for comparison against purified collagen standards, extant controls, and across various ancient specimens.
The above histological structures are generally isolated via demineralization using a dilute acid (Lindgren et al., 2018; Schweitzer, Wittmeyer, & Horner, 2007; Surmik et al., 2016); this allows their biomolecular histology to be investigated using a suite of molecular methods. Characterization of morphology for these structures has historically been accomplished using a combination of light microscopy and both of transmission and scanning electron microscopy (Armitage & Anderson, 2013; Lindgren et al., 2018; Schweitzer et al., 20052013; Schweitzer, Wittmeyer, & Horner, 2007; Surmik et al., 2016). Light microscopy is a practical method to rapidly screen specimens for the preservation of histological structures. The use of both transmission and scanning electron microscopy together is particularly advantageous. While both offer nanoscale optical resolution, the former images sample cross-sections while the latter sample surface (Bozzola & Russell, 1999; Handbook of Microscopy, 2019). Both methods are also readily capable of detecting a distinct ~67-nm banding pattern unique to collagen protein helices (Boatman et al., 2019; Gottardi et al., 2016; Lin et al., 1993; Tzaphlidou, 2005). Observation of this banding pattern indicates either the presence of a collagen helix or compounds replicating its structure.
Studying the chemical aspect of biomolecular histology generally requires localizing chemical signal to a specific histological structure. Two methods with precedence for use within molecular paleontology are ToF-SIMS and Raman spectroscopy:  
  1. ToF-SIMS rasters a micro/nanoscale-diameter ion beam in a square, grid-like pattern across a specimen's surface. At each point in the square analysis “grid,” the chemical content of the specimen's surface (uppermost ~1–2 nm) at that specific point is detected and recorded as a spectrum of molecular and fragment ions. A specific ion can then be plotted according to its recorded intensity at each point in the grid to form a molecular map that mirrors the area analyzed across the specimen's surface. The specific types of ions detected via this process vary depending upon specimen chemical makeup; this allows the unique histological structures of a specimen to be targeted so that chemical makeup can be connected to morphology (Sodhi, 2004; Thiel & Sjövall, 2011; Touboul & Brunelle, 2016). A few studies have employed ToF-SIMS to analyze ancient specimens (Lindgren et al., 2012201520172018; McNamara et al., 2016; Orlando et al., 2013; Schweitzer, Suo, et al., 2007; Surmik et al., 2016). One recent publication used the method to analyze the biomolecular histology of demineralized epidermis from an exceptionally preserved Jurassic ichthyosaur (Lindgren et al., 2018). Ionic fragments consistent with peptides or related compounds, along with polyaromatic hydrocarbons, were successfully localized to the ichthyosaur epidermis. Recorded intensities for polyaromatic hydrocarbon and peptide-related ion fragments (such as those detected in the Jurassic ichthyosaur (Lindgren et al., 2018)) can be compared across extant and ancient histological structures. For example, elevated levels of polyaromatic related ions in one specimen relative to another would be predicted to indicate a higher degree of chemical degradation (Buseck & Beyssac, 2014; Delarue et al., 2016; Oberlin, 1984; Sjövall et al., 2021; Thiel & Sjövall, 2011). This is one potential method for evaluating changes in fossil/subfossil biomolecular histology by geologic timepoint and depositional environment.
  2. Raman spectroscopy utilizes a monochromatic laser to irradiate (typically) a single point a few microns in diameter on a specimen surface (Ferraro et al., 2003; Pan et al., 2019; Smith & Dent, 2019). As the laser's photons contact the specimen surface, a small number of them are inelastically scattered by the specimen surface; that is, they either gain or lose energy after contacting the specimen surface (Raman & Krishnan, 1928; Smekal, 1923). The degree to which these photons change energy depends on the type of molecular bond vibration the photon interacted with within the specimen. Detecting the change in these photons' energies forms a spectrum revealing the types of molecular bond vibrations present where the laser contacted. This allows specific histological structures to be analyzed for the types of molecular bonds present in their chemical makeup (Ferraro et al., 2003; Hill et al., 2020; Pan et al., 2019; Smith & Dent, 2019). A recent study attempted to analyze the biomolecular histology of fossil tissues using Raman spectroscopy (Wiemann et al., 2018). However, perusal of their published findings raised questions as to whether some of their data represented true Raman signal or was an artifact of autofluorescence (Alleon et al., 2021). Raman spectroscopy with a laser wavelength below 250-nm is a well-established solution to eliminate autofluorescence (Abbey et al., 2017; Long et al., 2018) but has seen little use within molecular paleontology historically (Long et al., 2018). However, similar to the ion intensities with ToF-SIMS, Raman signal intensity for specific bond vibrations can be compared across extant and ancient specimen biomolecular histology. Indeed, this method has seen substantial use historically in correlating thermal history with molecular makeup for a wide range of humics and kerogen macromolecules in petroleum and soil science (Delarue et al., 2016; Ferralis et al., 2016; Schito et al., 2017).
Data collected using these described techniques can be correlated with the degree to which molecular sequences are recoverable from fossil and subfossil specimens. Both the intensity of Raman signal for specific bond vibrations and the relative ion abundances from ToF-SIMS can readily be compared against the degree to which a specimen preserves molecular sequence information. In the case of collagen peptides, both forms of electron microscopy can be used to evaluate the relative abundance of ~67-nm banding present within bone matrix. This too can be compared against the degree of type-1 collagen sequence information recoverable from a given specimen (similar to the preliminary case study described in Figure 1).

6 ADVANTAGES FOR SAMPLING OF ANCIENT SPECIMENS

Regarding fossil/subfossil specimen sampling, the above described methods generally function over a scale of micrometers to nanometers (Bozzola & Russell, 1999; Ferraro et al., 2003; Handbook of Microscopy, 2019; Marini et al., 2015; Pan et al., 2019; Smith & Dent, 2019; Sodhi, 2004; Thiel & Sjövall, 2011). Small samples of tens to hundreds of milligrams will suffice for any one of these molecular methods, provided care is taken during sample preparation. This limits the extent of destructive sampling necessary to study fossil/subfossil specimen biomolecular histology. Particularly, this allows for minimally destructive sampling of specimens that preserve exceptional morphology; this includes articulation, fossil organs, color, among other examples (Brown et al., 2017; Greenwalt et al., 2013; Lindgren et al., 201520172018; Manning et al., 2009; Yamagata et al., 2019). Examination of specimen biomolecular histology is hypothesized to yield insight into the general preservational state of such specimens at the molecular level. This would inform on whether future destructive molecular analyses, including sequencing, are justified for such morphologically exceptional specimens. If initial analysis of biomolecular histology suggests that a given “exceptional” specimen has limited potential for molecular sequence preservation, destructive sampling can be halted.
Furthermore, several recent studies have demonstrated isolated, disarticulate remains, even those stored for extended periods in museum collections, can often be used in molecular analyses in place of exceptionally preserved specimens that are more informative morphologically (Bertazzo et al., 2015; Cleland et al., 2016; Ngatia et al., 2019; Wiemann et al., 2018). The potential use of such specimens would improve stewardship of fossil and subfossil resources. Studying the biomolecular histology of such morphologically unexceptional specimens is hypothesized to further advance understanding on which geologic timepoints and depositional environments are most likely to harbor fossils/subfossils preserving ancient sequences. Advancing such knowledge, in this way, would then help limit the unnecessary sampling of more morphologically exceptional fossil/subfossil specimens that are otherwise unlikely to preserve sequenceable biomolecules at the molecular level, based on their diagenetic history.

7 CONCLUSION

Thermal setting and geologic age have been commonly used as proxies for predicting molecular sequence preservation potential (Demarchi et al., 2016; Hofreiter et al., 2015; Wadsworth et al., 2017; Welker et al., 2019). Late Pleistocene and Holocene specimens from cooler regions, especially permafrost deposits, have been shown to generally possess the highest preservation potential for molecular sequence information (Hofreiter et al., 2015; Letts & Shapiro, 2012; Ngatia et al., 2019; Wadsworth & Buckley, 2014; Welker et al., 2019). However, depositional environments are influenced by other variables including moisture (Briggs, 2003; Gupta, 2014; Lennartz et al., 2020; Lindahl, 1993; Schweitzer et al., 2019) and oxygen content (Briggs, 2003; Gupta, 2014; Lindahl, 1993; Schweitzer et al., 2019; Wiemann et al., 20182020), ion species present, and sediment composition (Briggs, 2003; Gupta, 2014; Lindahl, 1993; Schweitzer et al., 20142019). These confounding variables limit the usefulness of thermal setting and geologic age as proxies outside of a broad scale.
Direct analysis of fossil and subfossil biomolecular histology is a potential answer to this limitation. The biomolecular histology of a specimen's preserved cells and tissues reflects the cumulative effects of environmental variables upon its constituent biomolecules, including DNA and protein sequences (Briggs, 2003; Briggs et al., 2000; Gupta, 2014). Observed degradation of cell and tissue biomolecular histology is hypothesized to correlate with constituent biomolecules having undergone degradation. This agrees with the limited data in the primary literature on the correlation of biomolecular histology with sequence preservation potential (Asara et al., 2007; Rybczynski et al., 2013; Schweitzer et al., 2002; Schweitzer, Wittmeyer, & Horner, 2007). Thus, the preserved state of fossil/subfossil biomolecular histology is predicted to be an accurate proxy for molecular sequence preservation. A potential limitation to this approach is that some aspects of biomolecular histology may be beyond resolution or limit of detection for current molecular methods. However, modern molecular instrumentation regularly functions on the micro- and nanoscale in terms of resolution and limit of detection (Bozzola & Russell, 1999; Ferraro et al., 2003; Handbook of Microscopy, 2019; Marini et al., 2015; Pan et al., 2019; Smith & Dent, 2019; Sodhi, 2004; Thiel & Sjövall, 2011), thus minimizing this limitation as a potential obstacle. The use of fossil/subfossil biomolecular histology as a proxy for sequence preservation has potential to elucidate why ancient specimens of some formations and timepoints preserve sequences while others do not; such understanding would facilitate the selection of ancient specimens for use in future ancient DNA and paleoproteomic studies.

AUTHOR CONTRIBUTIONS

Landon A. Anderson: Conceptualization (lead); formal analysis (lead); investigation (lead); methodology (lead); project administration (lead); software (lead); supervision (lead); validation (lead); visualization (lead); writing – original draft (lead); writing – review and editing (lead).

Acknowledgements

The author thanks the journal editor-in-chief Dr. Gareth Jenkins, Ph.D., the associate editor Dr. Stefan Prost, Ph.D., and the anonymous reviewer for their time and efforts invested toward improving this manuscript. The author further thanks both his Ph.D. advisor, Dr. Mary H. Schweitzer, Ph.D., as well as Dr. Elena R. Schroeter, Ph.D., for their feedback and comments on various drafts of this manuscript. The author thanks Dr. Christopher L. Hill, Ph.D., for his assistance in providing and verifying information regarding specimens MOR 91.72, MOR 604, and MOR 605 used in this study. The author also thanks Dr. Lance C. Anderson, D.O., (degree expected Spring 2024) for his feedback which helped improve the manuscript's clarity and readability.
Additionally, several people and institutions contributed resources essential to the production of this manuscript in its finalized format. The author thanks the Museum of the Rockies for access to specimens used in this research: MOR 604 and 605, (donated to the Paleontology Department by Jerry and Kathy Doeden) and MOR 91.72 (Cultural History Department). The author thanks Miles Carson for his assistance and support in recovering and subsequently donating specimen YG 610.2397, and thanks the Yukon Paleontology Program of the Yukon Government for curating and permitting access to specimen YG 610.2397. The author also acknowledges that specimen YG 610.2397 was found on and recovered from the traditional territories of the Tr'ondek Hwech'in. The author additionally thanks the donors Lynn and Susan Packard Orr, and Vance and Gayle Mullis, for funding the data collection and article processing charge for this manuscript. Finally, this work was performed in part at the Chapel Hill Analytical and Nanofabrication Laboratory, CHANL, a member of the North Carolina Research Triangle Nanotechnology Network, RTNN, which is supported by the National Science Foundation, Grant ECCS-1542015, as part of the National Nanotechnology Coordinated Infrastructure, NNCI.

Competing Interests

There are no competing interests to declare.

Data Availability Statement

Data from this work is available in the Dryad public repository: https://doi.org/10.5061/dryad.8gtht76sq

References