Long-term exposure to ultraviolet (UV) light generates substantial damage, and in mammals, visual sensitivity to UV is restricted to short-lived diurnal rodents and certain marsupials. In humans, the cornea and lens absorb all UV-A and most of the terrestrial UV-B radiation, preventing the reactive and damaging shorter wavelengths from reaching the retina. This is not the case in certain species of long-lived diurnal birds, which possess UV-sensitive (UVS) visual pigments, maximally sensitive below 400 nm. The Order Psittaciformes contains some of the longest lived bird species, and the two species examined so far have been shown to possess UVS pigments. The objective of this study was to investigate the prevalence of UVS pigments across long-lived parrots, macaws and cockatoos, and therefore assess whether they need to cope with the accumulated effects of exposure to UV-A and UV-B over a long period of time. Sequences from the SWS1 opsin gene revealed that all 14 species investigated possess a key substitution that has been shown to determine a UVS pigment. Furthermore, in vitro regeneration data, and lens transparency, corroborate the molecular findings of UV sensitivity. Our findings thus support the claim that the Psittaciformes are the only avian Order in which UVS pigments are ubiquitous, and indicate that these long-lived birds have UV sensitivity, despite the risks of photodamage.
Photodamage arising from exposure to short wavelength light is a serious problem for numerous plants and animals, and in vertebrates, damage extends to the retina . In humans, the cornea and lens absorb all ultraviolet (UV)-A (315–400 nm) and most of the terrestrial UV-B (300–315 nm) radiation; the reactive and damaging shorter wavelengths of light are thereby prevented from reaching the retina. However, the screening of shorter wavelengths is not present in species that possess UV-sensitive (UVS) photoreceptors that permit vision into the UV. Among Eutherian mammals, UV sensitivity appears to be present only in mice and rats but is more widespread among marsupials [2,3]. Most are however nocturnal, so exposure to high light levels is relatively limited, and they are all relatively short-lived, thereby allowing less time for damage to accumulate. However, this is not the case in certain species of long-lived diurnal birds.
Birds are highly dependent on vision with a visual system that contains an elaborate mechanism for colour detection [4–6], with up to four classes of visual pigment present in single cones that most probably serve tetrachromatic colour vision [4–9]. Cone photoreceptors contain coloured oil droplets in the distal regions of their inner segments that are spectrally matched to the visual pigment and act to absorb all wavelengths shorter than a droplet-specific cut-off wavelength [10,11]; short wavelength absorption in the single cones arises from carotenoids, the concentration of which is resistant to dietary manipulation . In all species, the short wavelength-sensitive visual pigment is specified by the short wavelength-sensitive type 1 (SWS1) gene class; the ancestral form of this gene almost certainly specified a UVS pigment with peak sensitivity (λmax) below 390 nm . SWS1 pigments also exist in a violet sensitive (VS) form with λmax > 390 nm, and it is a VS version of the pigment that is present in most mammals including humans. At the base of the avian lineage, UV-sensitivity was almost certainly lost with the majority of extant species now possessing VS pigments . However, UVS pigments are found in 4 out of 14 avian Orders examined to date [13,14], where it has been re-acquired independently in each Order by a single amino acid substitution of Ser by Cys at site 90 [15–17]. In the four Orders containing species with UVS pigments (Passeriformes, Struthioniformes, Ciconiiformes, Psittaciformes), some species have VS rather than UVS pigments, but this is not found in the Psittaciformes where both the species (budgerigar; African grey) examined so far have UVS pigments. The molecular mechanism for the generation of UVS pigments differs in non-avian vertebrates , thereby reinforcing the conclusion that avian UVS pigments have a separate and more recent origin .
The use of UV waveband in reflectance-mediated cues in sexual selection, foraging and other visually guided behaviours has been extensively documented in birds [18–29]. There are, however, associated costs of UV sensitivity. UV light has a higher scattering, according to 1/λ4 (Rayleigh effect), causing chromatic aberrations that reduce spatial resolution and contrast, and hence detection of distant objects (reviewed in Bennett & Cuthill ). These physical obstacles can, nonetheless, be reduced by high-quality optical media and close distance to the object. More critically, UV light is damaging when absorbed by biological molecules, particularly DNA, and high levels of UV radiation are known to cause cellular damage to retinal and ocular tissue both in situ and in vitro, which mostly accumulates with age [31–34].
The presence of UVS cones in the avian retina was first demonstrated by microspectrophotometry in the Pekin robin, Leiothrix lutea  and subsequently by in situ hybridization in the budgerigar, Melopsittacus undulatus . The budgerigar is a member of the Order Psittaciformes that contains some of the most long-lived birds, such as the cockatoos and macaws . If UV-sensitivity is a conserved trait across the Order, this raises several issues regarding the use of UV signals, especially in relation to the effects on the retina of a life-time exposure to UV radiation. Preliminary evidence that UV-sensitivity is present has been obtained from a partial sequence of the SWS1 gene from the African grey parrot; Cys90 is encoded  as present in other avian species shown to possess UVS pigments [15,16]. The main objective of this study was, therefore, to assess whether UVS pigments are widespread throughout a diverse selection of long-lived parrot species (from Australia, South America, New Zealand and Indonesia) and to determine the transparency of the lens to UV light.
2. Material and methods
(a) cDNA synthesis and rapid amplification of cDNA ends
Poly(A)+ mRNA was extracted from Platycercus elegans adelaidae (rosella) retinae using the QuikPrep Micro mRNA Purification Kit (Amersham Biosciences, distributed by GE Healthcare Life Sciences, Little Chalfont, UK). Retinae had been stored in RNAlater (Ambion, Austin, TX, USA) since collection from free-living birds caught under licence in South Australia (34°57′ S, 138°49′ E). Single-stranded cDNA was synthesized using the 5′–3′ RACE Kit 2nd Generation (Roche Diagnostics, Burgess Hill, UK). Partial SWS1 opsin fragments were amplified using degenerate primer pairs BUD1F/BUD418R and AOASF/AOASR (electronic supplementary material, table S1). From these partial sequences, species-specific 3′- and 5′-end primers were designed to obtain the full-length coding sequence for the P. elegans SWS1 opsin gene (GenBank accession no. HM150794) using the 3′–5′ RACE Kit 2nd Generation (Roche Diagnostics).
(b) In vitro expression and regeneration of SWS1 opsin pigment
The PCR amplification of the coding sequence of the P. elegans SWS1 opsin and subsequent cloning into the pMT4 vector followed previously published protocols [38,39]. The PCR primers used are listed in electronic supplementary material, table S1. The expression construct containing the P. elegans SWS1 opsin was used to transiently transfect HEK 293T cells using GeneJuice Transfection Reagent (Novagen, distributed by Merck, Nottingham, UK). Tissue culture plates of 12 × 140 cm were used and at 48 h post-transfection, the cells were harvested and washed four times with phosphate-buffered saline (PBS-pH 7.0). Cell pellets were stored at −80°C prior to regeneration of the pigment. The regeneration, solubilization and purification of the pigments followed the method of Molday & MacKenzie  and have been briefly described in Carvalho et al. .
(c) Determination of λmax of expressed pigments
The absorbance spectrum of the purified pigment (dark spectrum) was determined by UV-vis spectroscopy using a Spectronic Unicam UV500 dual-beam spectrophotometer by scanning from 250 to 600 nm. The sample was then treated with hydrochloric acid (1 M) to denature the protein and re-scanned. This acid-treated spectrum was then subtracted from the dark spectrum to give a difference spectrum, which thereby avoids distortion caused by underlying absorbance and scatter by the protein. This was then used to determine the λmax by fitting a standard visual pigment template  using a Solver add-in to Microsoft Excel, which varies the λmax until the best fit to the template is found.
(d) Lens transmission analysis
The transmission of P. elegans lens was measured by scanning spectrophotometry from 300 to 700 nm .
(e) Amplification and partial sequencing of the SWS1 opsin gene
Retinal cDNAs from four endemic Australian parrot species were kindly donated by Prof. Lyn Beazley (University of Western Australia, Perth, Australia) and include the twenty eight parrot (Barnardius zonarius semitorquatus), sulphur-crested cockatoo (Cacatua galerita), Carnaby's black cockatoo (Calyptorhynchus latirostris) and the galah (Eolophus roseicapilla) (table 1). From the full-length sequence obtained for the P. elegans SWS1 opsin, gene-specific primers ROSELLA SWS1 F/ROSELLA SWS1 R (electronic supplementary material, table S1) were designed to the 5′- and 3′-end of the coding sequence. PCR amplification generated a product of around 1 kb for all four species of Australian parrots. These products were then cloned and fully sequenced. Genomic DNA for the remaining parrot species (table 1) was extracted from plucked blood feathers stored at −20°C. Partial SWS1 opsin gene sequences were obtained using primer pair BUD1F/BUD418R (electronic supplementary material, table S1), generating a fragment of around 1.8 kb encoding the complete sequence of exon 1 and intron 1 and a partial sequence of exon 2. PCR fragments were then cloned and fully sequenced.
(a) Platycercus elegans SWS1 opsin
The full-length coding sequence of the P. elegans SWS1 gene plus the 5′- and 3′-UTR regions has been deposited in GenBank (accession no. HM150794). Figure 1a shows the predicted amino acid sequence aligned with the budgerigar, pigeon (Columba livia) and chicken (Gallus gallus) SWS1 opsins. When compared with these other sequences, the P. elegans SWS1 sequence shows 94, 90 and 84 per cent nucleotide identity and 97, 90 and 84 per cent amino acid identity, respectively, to the budgerigar, pigeon and chicken SWS1 sequences. Figure 1b shows the phylogenetic relationships among the avian SWS1 opsin genes, placing the P. elegans sequence in close proximity to the budgerigar SWS1 gene and is consistent with the close evolutionary relationship of these two species.
(b) Spectral sensitivity of the Platycercus elegans SWS1 pigment
Inspection of the P. elegans sequence shows that the residue present at site 90 is a Cys (bovine opsin numbering), the same residue found in the budgerigar UVS SWS1 opsin (figure 1a) and shown by Wilkie et al.  to be responsible for UVS. In order to confirm that the P. elegans SWS1 pigment is also UVS, the full-length wild-type coding sequence was expressed in vitro; the resulting pigment yielded a λmax value in the UV at 363 nm (figure 2).
(c) Transmission of the Platycercus elegans lens
The spectrum of light transmission of the P. elegans lens is shown in figure 3. Two lenses were analysed and both showed around 70 per cent transmission at 350 nm, indicating that the P. elegans lens is transparent to UV light.
(d) Partial sequences of the SWS1 opsin gene
A further 13 species were studied as listed in table 1. These included representative species of South American macaws, Caribbean amazons, Indonesian and Australian cockatoos, and Australian broad-tailed parrots, plus the African grey parrot and the New Zealand Kea. Partial SWS1 opsin sequences of 393 bp of coding sequence that encompasses exons 1–2 were obtained from either cDNA or gDNA. The starting point for these partial fragments is at nucleotide +25 downstream of the translation start codon and extends to nucleotide position 417 in exon 2. The exon 1/2 boundary is located between nucleotide positions 346 and 347. Of the sequences obtained here, part of the African grey parrot SWS1 gene sequence has been previously reported (GenBank accession no. AY227186) by Odeen & Hastad , although this sequence extends only from nucleotide positions 219 to 278 in exon 1. The translated nucleotide alignment for these partial sequences is shown in figure 4. In all species, Cys90 is present and this is combined with Ala86 as found in the budgerigar SWS1 pigment .
The sequences of the SWS1 opsin gene presented here for a total of 14 Neotropical, Australasian and African species sampled across the Psittacidae, Cacatuidae and Nestoridae families, show that all have Cys90. Previous studies by Wilkie et al.  and Yokoyama et al.  have shown that in avian pigments, the presence of Cys90 confers UV sensitivity, whereas VS pigments from other species have Ser90. Furthermore, the in vitro-expressed rosella SWS1 pigment gave a peak at 363 nm, thereby confirming the presence of UVS pigments in the larger and longer lived parrot species. The Psittaciformes are therefore the only avian order where all species studied to date have a UVS pigment and is strong evidence that the Ser90Cys substitution responsible for this peak shift into the UV occurred at the base of the Psittaciformes lineage, and has been retained in all extant species.
Consistent with the presence of UVS pigments, the lens of P. elegans has a 50 per cent transparency cut-off at around 320 nm, which would allow a substantial proportion of UV light to reach the retina. This rises to 76 per cent transmission at the peak sensitivity of 363 nm of the expressed pigment. Coloured oil droplets that act as cut-off filters are routinely found in avian single cones. The cut-off wavelengths of these droplets are matched to the spectral sensitivity of the different cone pigments [12,46] and would therefore offer some protection to photoreceptor outer segments. UVS cones however have T-type (transparent) droplets that show no detectable absorbance from 330 to 800 nm, so these droplets would not significantly reduce the penetration of UV light into the outer segments of UVS photoreceptors. Nevertheless, all photoreceptor cells would be exposed to transmitted UV light proximal to the oil droplets; thus, the need for photoprotection for the nucleus and other organelles proximal to the oil droplets might be similar for all photoreceptor types, and indeed for many cell types in the retina. Whether long-lived birds compared with short-lived birds have increased lipopigment in the retinal pigment epithelium, or better mechanisms for removal of peroxidized lipids in the outer segments, remains to be determined.
The selective pressures that were responsible for the re-acquisition of UVS pigments in birds are unclear. The SWS1 opsin sequences for the 14 parrot species reported here are highly conserved: when compared with the budgerigar sequence, only a small number of nucleotide differences are observed, with very few of these differences being non-synonymous. A better understanding of the visual ecology of these birds will be essential in determining how UV sensitivity is employed. UV-reflecting plumages are found ubiquitously across all avian lineages, including parrots, suggesting that UV reflectance is an ancestral characteristic of the avian order [27,47,48]. The highest percentage (more than 10%) of UV reflectance is usually found in blue and white feathers, although feathers reflecting predominantly at longer wavelengths can have a biologically significant UV reflectance [28,47–49]. Parrots have the highest percentage of UV reflectance, especially in plumage patches when compared with all other avian orders, and in a survey of 143 species, all but three were shown to possess a considerable amount of UV-reflecting plumage .
With the exception of P. elegans that has an expected lifespan of around 15–20 years in captivity, all the other species investigated here are considered long-lived with lifespan of up to 50 years or more in captivity . UV radiation is known to damage several components of the eye such as the cornea, lens, the retinal pigment epithelium and the photoreceptors, leading to apoptosis in the cornea and pigment epithelium , decreased lens cell viability [50,51], the induction of oxidative stress  and the inactivation of anti-oxidant enzymes like catalase . UV radiation is also associated with several age-related conditions like cataract and macular degeneration [33,50]. In mammals, a long lifespan would appear to be associated with a loss of UV-sensitivity of the visual system with the peak sensitivities of the SWS1 pigments shifted to wavelengths greater than 400 nm . Only short-lived species such as the rat and mouse among the Eutheria have retained UVS pigments although, from the species studied so far, this may be more prevalent among the metatheria [3,11]. The penetration of UV radiation in the neural retina in long-lived parrot species might be expected therefore to cause irreversible damage [34,52] and raises the question as to how these species tolerate UV exposure. One explanation may be that birds show a lower rate of mitochondrial free-radical production in comparison with mammals [53,54]. Furthermore, cells of long-lived birds exhibit exceptional resistance to oxidative-induced damage when compared with shorter-lived birds and mammals , although the underlying genetic and molecular basis for this higher oxidative resistance is unknown. A third explanation could be the presence of carotenoids, which are present not only in the oil droplets of the retina, but also in the serum of parrots at concentrations comparable to that in other bird species [12,56]. Carotenoids are known to be effective antioxidants and immunostimulants in birds (reviewed in ), with lutein and zeaxanthin being especially effective in conferring protection against oxidative damage in the retina [58,59] and light-induced photoreceptor apoptosis . Similarly, psittacofulvin pigments, found so far only in parrot feathers, are thought to have considerable antioxidant properties [61–63], and could offer protection if also present in the retina.
If such processes are occurring in the retina, the combination of a lower production of free radicals and an increased resistance to oxidative damage could protect the eye against the damaging effects of UV light exposure in long-lived parrot species. A more detailed study of these potential protection mechanisms could provide valuable information on the survival of retinal cells. Such studies would reveal not only how extant long-lived diurnal terrestrial birds avoid UV-induced photodamage, but may also give new insights into UV protection, thereby facilitating improvements in human health associated with reducing or ameliorating UV-induced photodamage.
This research was conducted with the approval of the South Australian Wildlife Ethics Committee, the South Australian Department for Environment and Heritage, the Animal Ethics and Experimentation Committee of the University of Western Australia and the UK Home Office.
This work was supported by a PhD Strategic Studentship from the UK Biotechnology and Biological Research Council awarded to B.K., a Leverhulme Trust project grant awarded to D.M.H., a Leverhulme Fellowship to A.T.D.B. and a BBSRC project grant to A.T.D.B., M.L.B. and Dr K.L. Buchanan (Cardiff, Deakin). We are grateful to Prof. Ron Douglas, City University, London, for the lens transmission determinations; to Dr Rosalie Crouch, Storm Eye Institute, Medical University of South Carolina, USA, for the provision of 11-cis-retinal; to Dr Wayne Davies for technical help; and to David Paton, Leo Joseph, Ron Sinclair and Peter Bird for advice on P. elegans.
- Received May 23, 2010.
- Accepted July 7, 2010.
- This Journal is © 2010 The Royal Society