Analysis and comparison of protein secondary structures in the rachis of avian flight feathers

Feather specimens

All the feathers analyzed in this study were primary flight feathers, including domestic chicken (Gallus gallus domesticus), mallard duck (Anas platyrhynchos), African sacred ibis (Threskiornis aethiopicus), crested goshawk (Accipiter trivirgatus), collared scops owl (Otus lettia), budgerigar (Melopsittacus undulatus), and zebra finch (Taeniopygia guttata), to represent bird species with different flight modes. Primary flight feathers of domestic chickens and mallards were collected from an aviary at National Chung Hsing University. Primary flight feathers of budgerigar and zebra finch were collected from an aviary at National Tsing Hua University. Primary flight feathers of African sacred ibis, crested goshawk, and collared scops owl were plucked from the carcasses of wild individuals. The use of feathers from wild animals for research was approved by the Forestry Bureau of the Council of Agriculture of Taiwan (R.O.C.) under case no. 1091617349 (4/22/2020).

Paraffin-embedded feather sections

The sectioned feather samples from a serial 5-µm thick section of the paraffin embedding rachis of flight feathers were prepared for all species in this study. The rachis was soaked in liquid paraffin at temperatures in the range of 60–62 °C for 2 days. The paraffin-embedded rachis sectioned sample was transferred onto a Low-E slide (Kevley Technologies, Chesterfield, OH, USA), and embedded paraffin of the rachis sample section was removed by xylene washing before SR-FTIR microspectroscopic mapping.

EDC protein secondary structure prediction

Expression profiles in feathers and protein sequences of the epidermal differentiation complex (EPC) of chickens were annotated from Lin et al. (2021a) and Lin et al. (2021b). Secondary structures of EDC proteins were predicted using JPred4 (https://www.compbio.dundee.ac.uk/jpred/, accessed on 3 December 2021) (Drozdetskiy et al., 2015).

Attenuated total reflection Fourier transform infrared spectroscopy

The attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy analysis system uses a modulated infrared beam to enter an infrared transparent crystal with a high refractive index in the spectral range 4,000–650 cm⁻¹ at 45° of incidence. In this study, a high-pressure clamp is utilized to pressure the rachis medulla and dorsal cortex sample to make contact with the surface of the ATR crystal (PIKE MIRacle™, 3-reflection diamond/ZnSe crystal). Furthermore, the evanescence wave of the modulated infrared beam occurs on the contact surface of the ATR crystal; the depth of penetration (dp) of the propagating evanescence wave entering the sample ranges from approximately a hundred nanometers to several micrometers, as shown in the equation shown below; and the FTIR spectra of the feather samples were collected by using an FTIR spectrometer coupled with an LN-cooled MCT detector. ${\displaystyle dp=\frac{\lambda }{2\pi n_1\sqrt{sin\theta -{\left(n_1/n_2\right)}^2}}}$ where λ is the wavelength of incident mid-infrared light, θ is the incident angle of the crystal, and n ₁ and n ₂ are the refractive indices of the sample and ATR crystal, respectively. The FTIR spectra were acquired in the range of 4,000-650 cm⁻¹ and accumulated 128 scans at a spectral resolution of 4 cm⁻¹. The penetration (dp) of the infrared evanescent wave propagating into the sample was calculated to be approximately 2.15 µm at wavenumber 1,650 cm⁻¹ (the refractive index of the feather is approximately 1.55; the refractive index of the crystal is 2.4; the incident angle is 45°; and the wavelength is approximately 6.06 micrometers for wavenumber 1,650 cm⁻¹). Three samples of flight feathers of each avian species were analyzed.

Synchrotron-radiation-based Fourier transform infrared microspectroscopy

The spectral maps of the sectioned flight feather samples were acquired at the SR-FTIRM endstation of TLS 14A1 of the National Synchrotron Radiation Research Center (NSRRC) in Taiwan, including an FTIR spectrometer (Nicolet 6700, ThermoFisher Scientific, Madison, WI, USA) coupled with a confocal infrared microscope (Nicolet Continuum; ThermoFisher Scientific, Madison, WI, USA). The detailed procedures of SR-FTIRM mapping experiments are described in Lee et al. (2017). The confocal aperture of the confocal infrared microscope was set to 15 × 15 µm² and 10 × 10 µm² of step size for roast scanning of the sample area of interest. The Am I band is the main characteristic absorption of carbonyl groups of the peptide bond framework, showing lower vibration wavenumbers than those of chemicals with carbonyl functional groups due to the hyperconjugation of peptide bonds and forming various PSSs by intermolecular and intramolecular hydrogen bonds. The percentages of corresponding PSSs of feather samples in the Am I band spanning the spectral range of 1,720–1,580 cm⁻¹ were resolved by using the Peak Resolve function in OMNIC™. The automatic atmospheric suppression function in OMNIC™ was employed to eliminate the rovibrational absorption contributions from carbon dioxide and water in the ambient air. Three samples of flight feathers of each avian species were analyzed.

Spectral specificity of the cortex and medulla using attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy

The ratio of the absorption component band of the calculated PSS to the total area of Am I is shown in Fig. 3 (Figs. S2 & S3). In general, all medulla and cortex samples showed strong absorption attributed to PSS β-sheets, and β-keratins are well understood as major components of feathers. Additionally, the peak area percentage of approximately 10–15% attributed to characteristic absorption of the triple helix for collagen was resolved in the Am I band, consistent with the genomic expression of DNA related to collagens (Ng et al., 2015). Two characteristic absorption bands spanning the range of 1,641–1,623 cm⁻¹ and the range of 1,695–1,670 cm⁻¹ have been indicated as an antiparallel β-strand in a β-sheet structure for keratin (Zhang, Senak & Moore, 2011). Therefore, the peak-height ratio H_1,695/H_1,630 was suggested to be proportional to the relative distribution of the antiparallel β-strand component (1,695 cm⁻¹) to the parallel β-strand component (1,630 cm⁻¹) in a β-sheet structure (Chirgadze & Nevskaya, 1976). Previous studies of pelican and seagull feathers showed that the β-sheets primarily existed in the antiparallel conformation (Fraser & Suzuki, 1965). Collectively, the β-sheet structure was expected to be dominant in these feather samples. Although some epidermal differentiation complex (EDC) genes are expressed in feathers (Lin et al., 2021a; Lin et al., 2021b), most EDC members, especially those higly expressed in feathers, do not contain distinct secondary structures in predictions (Table S1).

The calculated peak area of the α-helix structure was approximately 9–12% in the Am I band for all feather samples; hence, we proposed that α-keratins might develop in mature flight feathers (Fig. 4). β-sheet structures were observed in the α-keratin filaments, which were suggested to increase the mechanical strength of feathers (Parry & North, 1998; Parry et al., 2002; Kreplak et al., 2004) to reduce stress during flight. Therefore, the content of α-keratins would be underestimated by the suggestion of only α-helix contribution in the structure of α-keratins. Based on our findings, the structure of β-sheets developed from α- and β-keratins plays an essential role in higher levels of assembly and in determining their mechanical properties (Fraser & Parry, 2009). The helical content was slightly higher in the flight feather cortex of chickens, ducks, and sacred ibis than in the medulla but was slightly lower in the flight feather cortex of goshawk, owl, budgerigar, and zebra finch (Table S2).

In addition, we also estimated proteins other than α- and β-keratins, in which distinct features are observed in the FTIR spectrum. The calculated contribution of the triple-helix structure in the Am I band was characteristic absorption for collagen type I; the other absorption bands were observed at 3,279 cm⁻¹, 3,052 cm⁻¹, 1,545 cm⁻¹, 1,292 cm^−1, and 1,260 cm⁻¹ and assigned to amide A, amide B, Am II, and amide III (Am III) bands, respectively (Fig. 2). In the study, the calculated results showed that the cortex of the flight feather might contain 11–16% collagen, whereas the medulla contained 13–16% collagen. The collagen content was slightly higher in the medulla than in the cortex, except for budgerigar and zebra finch.

Spectral map of PSSs of feather rachis sections

We extended the investigation of PSSs of proteins packed within the medulla by using lateral-resolved synchrotron-radiation-based Fourier transform infrared microspectroscopy (SR-FTIRM). The flight feather rachis samples of each species were sectioned by a series of continuous sectioning of the same rachis position, and herein, three cross-section samples were prepared and analyzed for each species. Lateral-resolved FTIR spectral images were constructed by using the peak height of the Am I band for the feather shaft medulla of seven species of birds, as shown in Fig. 5. The absorbance of the Am I band of the medulla for these birds varied among different regions, consistent with the suggestion of the varied density of packed cells in regions (Chang et al., 2019) (Figs. S4–S10), so that corneous protein was not evenly distributed in the medulla. The relative population of PSSs, such as α-helices, β-sheets, triple-helices, and random coil secondary structures, can be resolved by using Peak Resolve software in OMNIC. The analysis of the FTIR spectra of cross-section samples showed that a higher percentage of amino acids can be expected for α-helical peptides developed within the medulla cells of the flight feathers than has been reported and is not usually observed in β-sheet peptides (Ng et al., 2012) (Fig. S8).

α- and β-keratins of the medulla and cortex in different bird species

In this study, the accumulation of α- and β-keratins was observed in different parts of the rachis and ramus. α-Keratin has an important role in the early formation of rachises, barbs, and barbules, although feather- β-keratins are the major component of feathers. A deletion in two α-keratin genes (KRT75L4 and KRT6A (previously referred to as KRT75)) is associated with the frizzled feather phenotype (Ng et al., 2012; Dong et al., 2018). The rachis medulla with a defect in the KRT6A (previously known as the KRT75 gene) gene would cause the rigidity of feathers to decrease due to incomplete development and the rachis to bend (Ng et al., 2012). Despite this progress, few studies have revealed how the rachis becomes structured.

Cytoplasmic networks of intermediate filaments were built by obligate heteropolymers of types I and II α-keratins. The cornification process of claws, scales, beaks, and feathers of birds requires the deposition of β-keratins onto scaffolds of epidermal α-keratin filaments (Alibardi, 2013). Transcriptomic and evolutionary analyses were facilitated by manual annotation and curation of α- and β-keratin genes in avian genomes (Greenwold & Sawyer, 2010; Greenwold et al., 2014; Ng et al., 2014), suggesting that various combinations of α- and β-keratins contribute to morphological and structural diversity among avian epidermal appendages, with intrafeather differences largely attributed to differential expression of feather- β-keratins (Greenwold et al., 2014; Ng et al., 2014; Wu et al., 2015).

The flight feather cortex of ibis, goshawk, owl, budgerigar, and zebra finch contains higher helical content in the cortex of the flight feather than in the medulla compared to chickens and ducks who mainly live on the ground (Table S1). Goshawks and owls spend more time in woodlands and initiate mainly attacks on their prey from perches, whereas zebra finches and budgerigars are flappers with excellent maneuverability. Their flight feathers should endure stronger mechanical forces acting on the cortex in a short time. Higher contents of α-keratin may suggest that the mechanical properties provided by α-keratins in the cortex of flight feathers may have been previously disregarded. Further studies, including more avian species sampled from wide ranges of phylogenetic lineages, are required to perform a solid comparison of PSSs in birds of different flight modes.

Although it has been reported that unfavorable conditions such as nutrient-poor conditions in captivity during molting may affect the quality of avian feathers in some species of birds (Murphy, King & Lu, 1988; Dawson et al., 2000), we can eliminate this possibility because we did not observe that feathers from aviaries (chickens, ducks, budgerigar, and zebra finch) are significantly different from those derived from the wild (ibises, goshawks, and owls). Additionally, birds kept in aviaries are either domesticated birds or pet birds and should already be well adapted to artificial environments.

Collagen in the rachis

The FTIR spectra of the medulla of the feather rachis revealed crucial spectral evidence for collagen type I, as strongly supported by the characteristic absorption of the Am III band of protein in the range of 1,290–1,260 cm⁻¹, glycan residual (oligosaccharide residue) chemically attached to collagen in the range of 1,200–900 cm⁻¹, and triple-helix resolved at 1,637 cm⁻¹ in Am I band (Cooper & Knutson, 1995; Belbachir et al., 2009; Lee et al., 2017). Based on these findings, we strongly suggested that collagen type I develops in the medulla of the feather rachis.

Collagen genes are differentially expressed in regenerating feathers of different morphotypes (Ng et al., 2015), and collagens might play morphogenetic roles during embryonic feather development (Mauger et al., 1982). Type VI collagen is regulated by mesenchymal cell behaviors during early feather development in embryonic chickens (Kobayashi et al., 2005). The distribution of collagen types I, III, and V becomes heterogeneous during the formation of placode and disappears from the apex of growing feather buds, while their collagen fiber becomes denser in the rudiment (Sengel, 1990). Collagens have been proposed to have critical roles in the development of feather rudiments (Goetinck & Sekellick, 1972). However, the role of collagen in mature feathers needs further investigation, so it is interesting to determine if collagens play a mechanistic role in feathers.

Feather rachis morphology

The medulla of the rachis is composed of keratinocytes that are developed to grow into hollow, spongy, and porous structures (Sullivan et al., 2019). The types of keratinocytes also develop into different growth directions, pore sizes, and pore shapes; furthermore, different pore sizes and medullary arrangements are well understood in the feather shaft medulla for birds with different flight abilities and flight modes (Chang et al., 2019). The ratio of the diameter of the average pore size to the weight in the medulla of the sparrows that flutter quickly is relatively high, while the elongation and orientation angles of aligned cell bands in the medulla of the soaring flight differ from those of other species.

Based on our findings, a greater area of the hollow cell depletion zone was observed in the medulla of the cross-section feather shaft, which may enable the feathers to resist greater mechanical stresses during flight, especially for birds with a better flight ability (Wang & Meyers, 2017; Chang et al., 2019). In addition, the cortex of the shaft of sustained flyers also tends to be thicker on dorsal and ventral sides and thinner on lateral sides (Wang & Meyers, 2017; Wang & Sullivan, 2017; Chang et al., 2019). Accordingly, this may represent the difference in the medulla or cortical structure of the feather shafts of birds with various flight modes, giving them different flight abilities.