Characterization of visual pigments, oil droplets, lens and cornea in the whooping crane Grus americana

The assortment of photoreceptors in bird retinas is highly conserved across species. Most species of birds that have been investigated have retinas with four spectral classes of single cones, a set of double cones consisting of a principal and an accessory cell, and a rod photoreceptor class (Bowmaker, 1977; Bowmaker et al., 1997; Hart, 2004; Hart et al., 2000; Hunt et al., 2009; Wright and Bowmaker, 2001). Moreover, avian cone photoreceptors contain oil droplets within the distal region of the inner segment. These oil droplets are colored as a result of the presence of carotenoid pigments, which range in color from clear through pale green to yellow and red. These oil droplets act as long-pass filters, tuning the sensitivity of the underlying visual pigments to longer wavelengths, narrowing the sensitivity function of the underlying photoreceptor and reducing overlap with adjacent photoreceptor spectral types (Bowmaker, 1977; Bowmaker et al., 1997; Hart, 2004; Hart et al., 2000; Hunt et al., 2009).

Typically there are five different classes of oil droplet found in bird retinas, each of which is paired with a particular cone photoreceptor and visual pigment type (Table 1) (Bowmaker et al., 1997, Hart et al., 2000; Hart and Hunt, 2007; Hart, 2001b; Hunt et al., 2009). Although visual pigments are characterized by their wavelength of maximal absorbance (λ_max), oil droplets are classified by their cut-off wavelength (λ_cut) values, which describe the wavelength below which no significant light is transmitted (Lipetz, 1984). Avian rod photoreceptors that contain rhodopsin 1 (Rh1) visual pigments, do not have oil droplets. Long-wavelength sensitive (LWS) visual pigments are either found in single or double cones. LWS single cone oil droplets appear orange to red in color (‘R-type’). In double cones, the principal photoreceptor contains an oil droplet that appears colorless, pale green or yellow (‘P-type’) (Partridge, 1989; Hart, 2001a), whereas the accessory photoreceptor only occasionally contains a small droplet (Hunt et al., 2009). Cones with rhodopsin 2 (Rh2; cones expressing Rh2 opsin genes are designated as middle-wavelength sensitive or MWS) visual pigments contain yellow, ‘Y-type’, oil droplets. Cones with short-wavelength sensitive type 2 (SWS2) cone visual pigments have clear to pale green ‘C-type’ oil droplets that contain carotenoids absorbing strongly below 450 nm (Goldsmith et al., 1984; Hart, 2001a; Hart, 2001b; Partridge, 1989). Finally, short-wavelength sensitive type 1 (SWS1) visual pigments are expressed in cone photoreceptors that contain transparent, ‘T-type’, oil droplets that lack carotenoid pigments and have negligible absorbance across the ultraviolet and visible spectrum.

Many studies have investigated the components of avian visual systems, particularly the spectral characteristics of the visual pigments and oil droplets within the passerines (Hart, 2001b; Hart and Hunt, 2007; Hunt et al., 2009). Based on the type of SWS1 visual pigment expressed, avian visual systems can be classified as either violet sensitive (VS; λ_max=402–426 nm) or ultraviolet sensitive (UVS; λ_max=355–380 nm). To understand the potential functional differences of these two broad visual categories, characterization of additional visual system components are needed. Most significant are the transmission characteristics of the lens and cornea, which can effectively absorb UV radiation and define the limits of photoreceptor UV sensitivity. In relation to UV vision, Hart (Hart, 2001b) hypothesized that smaller birds are more likely to have UVS pigments, possibly either to improve locomotive flight ability or because the characteristics of unpigmented ocular media allow transmission of enough UV to be useful for vision only in small eyes. Lind et al. (Lind et al., 2014) found a strong correlation between eye axial length (a proxy for body size) and ocular media transmission of 24 species of bird, supporting the hypothesis that eye size constrains avian sensitivity to UV.

Of particular interest are the effects of UV radiation on visual system capabilities. UV radiation tends to be more strongly scattered than visible light, reducing contrast. It also is refracted more strongly than visible light by the cornea and lens, and the resulting longitudinal chromatic aberration could further reduce spatial resolution and contrast (Bennett and Cuthill, 1994). Furthermore, UV wavelengths of radiation can be damaging to retinal tissues, with accumulating effects that have the potential to be severe for long-lived species (Sliney, 2002; Zuclich, 1989). However, many larger, long-lived birds have UV-sensitive visual systems, including UV-transmissive ocular media (Carvalho et al., 2011). The question of how large, long-lived birds balance the ecological demands leading to the evolution of UVS with the potential for UV-induced retinal damage remains largely unanswered.

In this study, we characterize the visual system of the whooping crane Grus americana (Linnaeus 1758). This study represents the first characterization of the visual system for a species from within the Gruiiformes, which includes measuring visual pigment and oil droplet absorbance, as well as examining the lens and cornea for UV absorbance properties. Whooping cranes are large birds that can live up to 40 years in captivity (Wasser and Sherman, 2010). Birds with larger eyes, such as G. americana, are predicted to use a violet-sensitive SWS1 visual pigment to improve image quality by decreasing defocus due to longitudinal chromatic aberration (Hart and Hunt, 2007). Additionally, ocular media that absorb a large percentage of UV radiation would reduce the retinal damage due to ultraviolet light over the long life-span of this species. As a further protection against retinal damage, corneal and lens materials are predicted to absorb UV radiation in large, long-lived birds. Because these birds have distinct visual behaviors (Cronin et al., 2007) and are subject to intensive conservation efforts, characterization of crane vision contributes to our knowledge of the ecology of this species for management purposes and more broadly provides insight to the ecological and evolutionary forces shaping avian visual systems.

Five full-length opsin transcript sequences were isolated from whooping crane retinal mRNA. Based upon phylogenetic analysis, these transcripts are related to other avian opsin sequences and correspond to the five major vertebrate visual pigment classes: RH1 (1056 bp), RH2 (1068 bp), SWS1 (1044 bp), SWS2 (1029 bp) and LWS (1095 bp) (Fig. 1). Using in vitro expression studies, we found that these five opsins form visual pigments that fall within the known variation for bird rod and cone photoreceptors (Fig. 1; Table 1). The four cone visual pigments had peak absorbance (λ_max) values of 404 nm (SWS1), 450 nm (SWS2), 499 nm (RH2) and 561 nm (LWS) (Fig. 1). The rod visual pigment had a measured λ_max of 500 nm (RH1). Based on the λ_max of the G. americana SWS1 visual pigment (404 nm), the whooping crane visual system can be classified as violet sensitive (VS). Similar to other birds, violet sensitivity in the whooping crane SWS1 visual pigment corresponds to the presence of amino acid residues Cys86 and Ser90.

We measured the spectral absorbance of 36 oil droplets from a single whooping crane retina. Based on color, as well as the shape and calculated λ_cut of the measured absorbance curves, five types of oil droplets were characterized in the G. americana retina with λ_cut values of 576 nm (R-type), 522 nm (putative Y-type), 506 nm (putative P-type), 448 nm (C-type) and negligible absorptance across all wavelengths (T-type) (Table 2, Fig. 2). The calculated λ_cut values for each oil droplet type either fell within values reported for other avian species (R-type, C-type, T-type), or were slightly red-shifted relative to previously reported values (P- and Y-type) (Tables 1, 2). The absorptances of the whooping crane cornea and partial lens tissue were both low (Fig. 3). Relevant to the question of G. americana UV sensitivity, both tissues had similar rises in absorptance in the UV range starting at ~420 nm.

Five classes of opsin mRNA sequences were isolated from the whooping crane retina. Expression of these opsin transcripts demonstrates that whooping cranes have a typical bird visual system, containing four visual pigments in cone photoreceptors and one visual pigment expressed in rod photoreceptors. The measured λ_max from heterologous expression of all of the crane visual pigments fall within known avian absorbance peak variation for each spectral class. Of particular interest is the λ_max of the SWS1 visual pigment, where a 404 nm peak absorbance gives the whooping crane a violet-sensitive visual system.

Residue 90 is the main amino acid controlling VS/UVS in the SWS1 visual pigments of birds (Wilkie et al., 2000; Yokoyama et al., 2000). Based on ancestral state reconstructions of the three amino acid sites (86, 90 and 93), which determine whether SWS1 visual pigments are either UVS or VS, avian color vision has shifted between VS and UVS at least 14 times (Ödeen and Håstad 2013). Additionally, violet sensitivity has been hypothesized to be the ancestral state for bird SWS1 visual pigments (Hunt et al., 2009), although this hypothesis is tenuous because only one of the three sites can be reconstructed unequivocally (Ödeen and Håstad 2013). Residue 86 has also been shown to play a role in shifts between VS and UVS visual pigments in vertebrates. In previous studies of the spectral sensitivity of bird SWS1 visual pigments, site-directed mutagenesis demonstrated that both Ser86Cys and Ser90Cys mutations generate a UV shift in λ_max (Shi and Yokoyama, 2003; Wilkie et al., 2000). Based on sequence data alone, the gene encoding SWS1 from two species from within the Gruiformes – the crowned crane Balearica pavonina (Gruidae) and the common coot Fulica atra (Rallidae) – are both predicted to have a λ_max value of 406 nm based upon amino acid substitution of the Cys86 residue (Ödeen and Håstad, 2003). Our study represents the first spectral characterization of any crane visual pigments to test these predictions. The whooping crane SWS1 gene sequence contains the same Ser90 and Cys86 amino acid residues as characterized in other Gruiformes, and the 404 nm λ_max matches the predicted values for the crowned crane and common coot extremely well (Ödeen and Håstad, 2003).

Avian visual systems generally contain five types of oil droplet associated with specific photoreceptor types and we characterized these five types of oil droplet in the whooping crane. Three types can be clearly classified as the T-type, C-type and R-type droplets (Table 2). Of these, the C-type and R-type oil droplets exhibit typical avian absorptance curves. However, when compared with data from other birds, the T-type droplet appears to have a relatively high absorptance (~0.2), particularly below 400 nm (Fig. 2). It is possible that the high absorptance of the T-type droplet is a measurement artifact due to the microspectrophotometry beam (2 μm), which is slightly larger than the diameter of the droplets being measured. If this high absorptance is real, it could indicate there are low levels of carotenoid in the T-type droplet that provide more of a filtering effect compared with other bird visual systems. The remaining two types represent the Y-type and P-type droplets, which can have very similar λ_cut values. Because we were unable to use fluorescence to help unambiguously classify these oil droplet types by spectral characterization (Hart, 2001a), the Y- and P-type λ_cut values are putative. Based on λ_cut ranges from other avian species and on the function of oil droplets as long-pass filters, we putatively assigned the droplets with λ_cut values of 522 nm as Y-type, and the droplets with 506 nm λ_cut values as P-type droplets (Table 2; Fig. 4).

The gruiform species that have been studied previously have unique Y-type and P-type oil droplet properties. In general, gruiforms have more intensely pigmented (i.e. more orange) Y-type droplets than those found in most other birds (Hart, 2001a). Although the spectral transmittance characteristics of single cone oil droplets tend to vary little across the retina, in the dusky moorhen Gallinula tenebrosa (Gruiformes, Rallidae) the P-type droplets vary in color across the retina, from almost as orange as the Y-type oil droplets to a pale yellow (Hart, 2001a). The whooping crane has P-type and Y-type oil droplets that have similar absorbance characteristics based on our putatively assigned λ_cut values. Additionally, the most abundant oil droplet type across an avian retina is the P-type, which on average represents 29.0–55.5% of the photoreceptors (Hart, 2001a). The dusky moorhen (Gallinula tenebrosa; Gruiformes) retina is composed of 20.2% Y-Type and 33.0% P-Type oil droplets (Hart, 2001a). Based on the absorbance similarities of our Y- and P-type droplet classes (Table 2), P-type λ_cut variations across the retina and differences in the composition of retinal oil droplets, it is also possible that our putatively assigned droplet types represent either P-type droplets from different retinal regions or mixtures of Y-type and P-type droplets. Because of the limitations of working with an endangered species, particularly with acquiring fresh retinal tissues, our sampling of oil droplet types in the whooping crane was limited. If material becomes available, future studies should characterize the yellow and pale oil droplet more thoroughly, and look for variations in λ_cut values in different retinal regions.

The T-type oil droplet class has negligible absorbance, therefore the sensitivity of the SWS1 photoreceptor is determined by the λ_max of the visual pigment and the absorbance of the ocular media (Hart, 2001b). This is similar to the findings of Lind et al. (Lind et al., 2014), who modeled the visual system of 38 bird species based on visual pigment λ_max and ocular media transmittance. This study found that color discrimination in bright light is mostly dependent on the visual pigment (UVS or VS), regardless of the characteristics of the ocular media, suggesting that ocular media spectral tuning is mainly relevant for detecting weak UV signals. At the low end of VS bird vision, SWS1 sensitivity overlaps with the upper part of the UV range (e.g. 380 nm), as is the case with the 404 nm SWS1 whooping crane visual pigment. Consequently, limited UV sensitivity may be possible despite the properties of the cornea and lens, given that whooping cranes are mostly white birds that are active during the day when reflectance would be maximized in the UV portion of the spectrum.

Birds with VS visual systems tend to have ocular media with longer-wavelength transmission (measured as λ_T0.5, the wavelength at half the measured transmission maximum) values, ranging from 335–378 nm, whereas those with UVS have values ranging from 316–343 nm (Hart and Hunt, 2007). Much of the ocular media absorbance occurs in the lens, with the cornea contributing little to overall UV absorbance, therefore our corneal absorbance data are not surprising. The absorbance measured in the whooping crane lens is more surprising, particularly in the UV portion of the spectrum (Fig. 3). Because our measurements were made on pieces of dissected lens tissue, we have not fully characterized the intact lens UV absorbance characteristics and we predict that intact whooping crane lenses probably absorb much more UV. The drop in absorptance below ~375 nm may be due to leaching of some of the lens pigments from lens pieces, or due to some uncharacterized fluorescence corrupting scans. Additionally, many birds have multifocal lenses (Lind et al., 2008), and although the focal properties within gruiform species are unknown, our measurements may have missed lens areas important to UV absorbance. Despite these potential artifacts, the rise in absorptance in the blue portion of the lens spectrum suggests that the whooping crane intact lens most likely has a λ_T0.5 value somewhere in the 380–410 nm range, consistent with other birds with VS visual systems.

Similar to the whooping crane, the visual systems of several other large birds have been characterized as VS, including the turkey (Meleagris gallopavo) (Hart et al., 1999), peafowl (Pavo cristatus) (Hart, 2002) and ostrich (Struthio camelus) (Wright and Bowmaker, 2001). Other larger bird species with violet-sensitive visual systems have expected λ_T0.5 values in the longer-wavelength ranges of UV light, between 358 and 370 nm (Hart et al., 1999; Hart, 2002; Wright and Bowmaker, 2001). Interestingly, of those species with SWS1 visual pigment λ_max values that are similar to those we found in the whooping crane, there is a large variation in ocular media transmission (Ostrich, SWS1 λ_max=405 nm, ocular media λ_T0.5=377 nm; wedge-tailed shearwater, SWS1 λ_max=406 nm, ocular media λ_T0.5=335 nm) (Hart and Hunt, 2007). One possibility for this variation is related to eye size; the wedge-tailed shearwater eye is smaller than the ostrich, reducing chromatic aberration and thus permitting an ocular media with a shorter λ_T0.5. However, the variation in ocular media transmission in relatively large eyes also demonstrates that long optical paths do not necessarily absorb the UV portion of the spectrum. This suggests that there may be some fine-scale tuning of visual system UV sensitivity in the absorbance properties of the ocular media based on the visual ecology of the species. Despite not having fully characterized lens UV absorbance properties, the possibility exists of limited UV vision, particularly in bright light environments, in the whooping crane based on a SWS1 visual pigment shifted towards UV wavelengths. This suggests there may be a role for functional UV sensitivity in whooping crane visual ecology.

Retinal damage due to long-term UV exposure is a potential issue for long-lived birds such as the whooping crane. Therefore, to avoid retinal damage, long-lived birds might be expected to have UV-absorbing ocular media, making UV-sensitive photoreceptors superfluous (Carvalho et al., 2011). Additionally, studies of other avian visual systems found a strong correlation between eye size and the ocular media transmittance, with larger eyes having longer λ_T0.5, the wavelength at half the measured transmission maximum (Lind et al., 2014). The whooping crane visual system in part matches these predictions by having a violet-sensitive SWS1 visual pigment.

Within birds with VS visual pigments there is variation in ocular media transmittance (Lind et al., 2014). If whooping crane lenses have low UV absorbance, mechanisms to repair UV damage would be needed, as with other long-lived birds having UV-sensitive visual systems (e.g. Psittaciformes) (Carvalho et al., 2011). It is also possible that the extremely violet-sensitive SWS1 visual pigment is a compromise between the longitudinal chromatic aberration that would result from transmission of short-wavelength UV light (necessary to excite UVS pigments) and the need to detect UV signals, perhaps in plumage. Nearly all white feathers reflect significant amounts of UV (Eaton and Lanyon, 2003; Mullen and Pohland, 2008) and the ability to detect some UV might contribute to the overall brightness of the reflected light from the white G. americana feathers. The violet-sensitive G. americana visual system, with a SWS1 absorbance peak at 404 nm, would be capable of detecting signals in the UV range.

In summary, we measured the spectral characteristics of the Grus americana visual pigments, oil droplets, cornea and partial lens tissue. The whooping crane contains a typical set of avian visual pigments and oil droplets, and as predicted, a violet-sensitive visual system. Further work characterizing the distribution of oil droplets, as well as the potential for variation in oil droplet absorbance, across the retina would further contribute to the overall understanding of whooping crane visual systems. Measurements of intact corneal absorbance indicate this tissue does not contribute to UV light filtering in the whooping crane eye, whereas our measurements of lens pieces suggests the presence of some near-UV absorbance in the lens as a whole. The presence of a short SWS1 visual pigment in a diurnal animal suggests that whooping cranes have a visual system capable of detecting UV radiation, warranting future studies of whooping crane plumage reflectance. The difference in the absorbance characteristics of the ocular media between species with VS visual systems suggests that rather than tuning either visual pigments or oil droplets, some birds may fine-tune UV sensitivity with ocular media absorbance.

Tissues were collected from a single captive individual of the whooping crane Grus americana from a breeding colony established at the USGS Patuxent Wildlife Research Center, MD, USA. The animal was euthanized because of circumstances unrelated to this study, and eye and muscle tissues were collected immediately after euthanization. Because the whooping crane is an endangered species and tissue from additional individuals was not available, retinal tissue from one eye was used to determine the spectral characteristics of the oil droplets, cornea and lens, whereas the other was used for molecular techniques to isolate and characterize the opsin genes. Collected tissues were dissected from the bird and either placed on ice until use (for spectral measurements) or immersed in either ethanol (muscle tissue) or RNAlater (retinal tissue; Qiagen, Valencia, CA, USA) and stored at −80°C for use in DNA or RNA extractions for molecular studies.

Retinal tissue was dissected out of a single, fresh crane eye and used in standard RNA extraction protocols (TRIzol, Life Technologies, Carlsbad, CA, USA). Retinal total RNA was used for first strand synthesis with an oligo(dT) primer and Superscript III reverse transcriptase (Life Technologies, Carlsbad, CA, USA). The resulting cDNA was used for 3′ RACE procedures with degenerate primers based on alignments of published bird and squamate opsin sequences for each class of vertebrate visual pigments (RH1, RH2, LWS, SWS1, SWS2; Table 3). Positive double-stranded RT-PCR products were cloned using the pGEM-T Easy vector system (Promega, Madison, WI, USA), and individual clones screened by sequencing with vector primers. Once a portion of each gene was isolated, if the 5′ end of the transcript was missing 5′ RACE was performed using 3′ gene-specific primers using the 5′/3′ RACE kit, 2nd generation (Roche, Madison, WI, USA). Full-length transcripts were confirmed by sequencing with a proofreading reverse transcriptase (AccuScript Hi-Fi Reverse Transcriptase, Agilent Technologies, Santa Clara, CA, USA) and polymerase (PrimeSTAR HS DNA polymerase, Takara, Otsu, Japan) using gene-specific primers designed against the 5′ and 3′ ends of each transcript (Table 3). All PCR products were sequenced by Genewiz (Germantown, MD, USA). Full-length opsin transcript sequences have been deposited to the NCBI GenBank database (RH1 – KM508488; RH2 – KM508489; LWS – KM508487; SWS1 – KM508490; SWS2 – KM508491).

To place the characterized opsin sequences in an evolutionary context, whooping crane transcripts were aligned with publicly available opsin sequences from birds including Corvus macrorhyncos (RH1, BAJ05379), Serinus canaria (RH1, CAB91997; RH2, CAB91995; SWS1, CAB91993; SWS2, CAB91994; LWS, CAB91996), Taeniopygia guttata (RH1, AAF63461; RH2, NP_001070164; SWS1, AAF63463; SWS2, NP_001070165; LWS, NP_001070170), Columba livia (RH1, AAD32241; RH2, AAD32242; SWS1, AAD38034; SWS2, AAD38035; LWS, AAD38036), Gallus gallus (RH1, P22328; RH2, P28683; SWS1, P28684; SWS2, NP_990848; LWS, NP_990771), Melopsittacus undulatus (RH1, AAC41247; RH2, AAC41246; SWS1, CAA72483), Anas platyrhynchos (RH1, AAC41245; RH2, EOA96196), Ptilonorhynchus violaceus (SWS2, AFK82745; LWS, AFK82727), Chlamydera nuchalis (SWS2, AFK82748; LWS, AFK82730), Chlamydera maculata (SWS2, AFK82747; LWS, AFK82729), Ailuroedus crassirostris (SWS2, AFK82743; LWS, AFK82725), Scenopoeetes dentirostris (SWS2, AFK82744; LWS, AFK82726), Calyptorhynchus latirostris (SWS1, ADG96042), Cacatua galerita (SWS1, ADG96044), Eolophus roseicapillus (SWS1, ADG96043), Barnardius semitorquatus (SWS1, ADG96041), Platycercus elegans (SWS1, ADG96036), Cyanistes caeruleus (SWS1, AAP30082), Palacrocorax carbo (SWS1, ABS86975), Spheniscus humboldti (SWS1, CAC20913) and Sericulus chrysocephalus (LWS, AFK82728). Amino acid sequences were aligned using MAFFT (Katoh et al., 2005; Katoh et al., 2002) and the resulting alignment used to estimate phylogenetic relationships and node confidence as bootstrap values using RAxML (Stamatakis, 2006; Stamatakis et al., 2008). Based on previously published phylogenies of opsin relationships (Porter et al., 2012), the phylogeny was rooted using avian (Gallus gallus, ACX32474; Taeniopygia guttata, NP_001266194), anuran (Xenopus laevis, NP_001165363) and squamate (Anolis carolinensis, ACX32472) sequences from the clade designated ‘vertebrate ancient (VA) opsins’, which are evolutionarily basal to the vertebrate visual pigments (Philp et al., 2000; Porter et al., 2012).

To characterize visual pigment spectral absorbance in vitro, each full-length opsin gene was first cloned into the mammalian expression vector pMT3 and appended with the last 15 amino acids of bovine rhodopsin (Franke et al., 1988; Oprian et al., 1987). Because we initially had difficulty characterizing the 5′ end of the LWS opsin transcript (~30 bp), a construct consisting of the first 16 bp from the Columbia livia LWS opsin followed by the near-full length whooping crane LWS opsin was created for in vitro expression. Based on the known function of the opsin N-terminal region in proper protein localization but not in spectral tuning, this construct allowed us to characterize the whooping crane LWS visual pigment absorbance. All constructs were expressed transiently in HEK 293A cells using Turbofect transfection reagent (Thermo Scientific, Waltham, MA, USA). Briefly, 10 μg of plasmid DNA was mixed with 20 μl Turbofect and incubated in 1.0 ml Dulbecco's Modified Eagle Medium (Life Technologies, Carlsbad, CA, USA) for 20 min and added to an 80% confluent 100×20 mm plate of HEK 293A cells. Cells were harvested 48 h post transfection. Protein expression was confirmed by western Blot analysis using anti-1D4 primary antibody and an alkaline-phosphatase-conjugated secondary antibody (data not shown). Blots were visualized after incubation with AttoPhos fluorescent substrate (Promega, Madison, WI, USA) on a Storm 840 molecular imager (GE Healthcare Life Sciences, Pittsburgh, PA, USA).

After confirmation that each construct expressed opsin in the HEK 293A cells at sufficient levels, proteins were purified for spectral characterization. All purification steps were performed under dim red light at 4°C. Thirty 100×20 mm plates of transiently transfected HEK 293A cells were harvested and washed once in 1× PBS. Before solubilization, photopigments were reconstituted in 5 ml of 10 mmol l⁻¹ MES/150 mmol l⁻¹ NaCl buffer (pH 6.0) containing 40 μmol l⁻¹ 11-cis retinal for 1 h. Membranes were solubilized in 10 mmol l⁻¹ MES in 150 mmol l⁻¹ NaCl buffer (pH 6.0), 1% n-Dodecyl-β-D-maltoside (DM), 0.2 mg ml⁻¹ PMSF for 2.5 h and spun in a clinical centrifuge to pellet debris. Homogeneous protein supernatant was incubated with anti-1D4 Sepharose column matrix for 16 h (Wilkie et al., 2000). The photopigment/column matrix complex was washed 10 times with 10 mmol l⁻¹ MES in 150 mmol l⁻¹ NaCl buffer (pH 6.0), 0.1% DM to remove excess chromophore. Photopigment was eluted from column matrix in 0.1% DM, 80 μM 1D4 peptide. Eluates were concentrated to ~250 μl using Amicon Ultra 30K MWCO centrifugal filters (Millipore, Billerica, MA, USA). Spectrophotometric measurements were taken from 240 to 700 nm in 1 nm increments with a Hitachi U3300 UV-Vis Spectrophotometer.

Microspectrophotometry was used to measure the spectral absorbance of photoreceptor oil droplets and to investigate the UV absorbance of intact cornea and partial lens tissue of G. americana eyes. All tissues for absorbance measurements were collected from a single eye. For all measurements (lens, cornea and oil droplets), a 2 μm beam was placed in the tissue of interest (oil droplet, lens or cornea) and absorbance was measured at 1 nm intervals from 350 to 700 nm. For oil droplet measurements, multiple droplets of each class, as determined by color to the human eye, were measured from a single retina (Table 2). The best oil droplet absorbance spectra for each color class (n=3–9) were converted to absorptance, averaged and used to calculate the λ_cut values for each drop type, defined as the wavelength of the intercept at the value of maximum measured absorptance by a line tangent to the absorptance curve at half-maximum measured absorptance (Lipetz, 1984). Although the best way to characterize ocular media transmission is in situ measurement of lens, cornea and vitreous humor spectral properties (Lind et al., 2014), the opportunistic availability of the whooping crane tissue and limited sample size (i.e. one animal, two eyes) made it impossible to prepare the materials required for this technique before the tissue became unusable. Therefore, we used microspectrophotometry to investigate the absorbance properties of the isolated cornea and pieces of the lens tissue, as a measure of whether or not whooping crane ocular media absorbed significant amounts of UV radiation. For these measurements, the cornea and lens were dissected from a single eye. Corneal measurements were made on whole cornea tissue sandwiched in 1× PBS between two UV-transmissive coverslips. For lens measurements, pieces of lens tissue were used in the same set-up (i.e. in 1× PBS held in place between UV-transmissive coverslips). For both lens and corneal measurements, the best absorbance measurements (n=4 and 7, respectively) were averaged and converted to absorptance. Typically, lens and corneal spectral characteristics are described in terms of ocular transmission [λ_T0.5, the wavelength at half the measured transmission maximum (Emmerton et al., 1980)]. However, given the methods used to characterize absorbance and the low UV absorbance measured in both tissue types, calculating λ_T0.5 was not possible.

The authors would like to thank Laura Seidman, Ifeolu Akinnola, and Elelbin Ortiz for laboratory assistance, P. Mullen for a helpful discussion of UV reflection in the Gruiformes, and two anonymous reviewers for valuable comments to improve the manuscript.