, Volume 91, Issue 12, pp 585–588

Retinal cryptochrome in a migratory passerine bird: a possible transducer for the avian magnetic compass


  • Andrea Möller
    • Zoologisches Institut, Fachbereich Biologie und InformatikJ.W. Goethe Universität
  • Sven Sagasser
    • Institut für Tierökologie und ZellbiologieTierärztliche Hochschule Hannover
    • Zoologisches Institut, Fachbereich Biologie und InformatikJ.W. Goethe Universität
  • Bernd Schierwater
    • Institut für Tierökologie und ZellbiologieTierärztliche Hochschule Hannover
Short Communication

DOI: 10.1007/s00114-004-0578-9

Cite this article as:
Möller, A., Sagasser, S., Wiltschko, W. et al. Naturwissenschaften (2004) 91: 585. doi:10.1007/s00114-004-0578-9


The currently discussed model of magnetoreception in birds proposes that the direction of the magnetic field is perceived by radical-pair processes in specialized photoreceptors, with cryptochromes suggested as potential candidate molecules mediating magnetic compass information. Behavioral studies have shown that magnetic compass orientation takes place in the eye and requires light from the blue-green part of the spectrum. Cryptochromes are known to absorb in the same spectral range. Because of this we searched for cryptochrome (CRY) in the retina of European robins, Erithacus rubecula, passerine birds that migrate at night. Here, we report three individually expressed cryptochromes, eCRY1a, eCRY1b, and eCRY2. While eCRY1a and eCRY2 are similar to the cryptochromes found in the retina of the domestic chicken, eCRY1b has a unique carboxy (C)-terminal. In light of the ‘radical-pair’ model, our findings support a potential role of cryptochromes as transducers for the perception of magnetic compass information in birds.


Magnetic orientation is widespread among animals (R. Wiltschko and W. Wiltschko 1995), yet the processes forming the sensory basis of magnetoreception remain largely unknown. By using behavioral tests with migratory orientation as a criterion, the functional mode of the magnetic compass has been analyzed in detail in small night-migrating passerines; for example in European robins, Erithacus rubecula (R. Wiltschko and W. Wiltschko 1995; W. Wiltschko and R. Wiltschko 2002; W. Wiltschko et al. 2004). Normal orientation was observed under monochromatic blue, turquoise, and green light (W. Wiltschko and R. Wiltschko 2001; Muheim et al. 2002). This wavelength dependency suggests a possible involvement of photoreceptors in magnetoreception.

Theoretical models of mechanisms that could provide birds with magnetic compass information assume a crucial role of specialized photopigments (Leask 1977; Schulten and Windemuth 1986). The currently favored ‘radical-pair’ model (Ritz et al. 2000) proposes that macromolecules are raised by photon absorption to singlet-excited states and subsequently generate radical pairs. Singlet pairs may be converted into triplet pairs, with the triplet yield depending on the alignment of the molecules in the ambient magnetic field (for details see Ritz et al. 2000). By comparing the triplet yield in the various spatial directions, birds could obtain directional information from the magnetic field. Magnetoreception by ‘radical-pair’ mechanisms thus requires receptors aligned in all spatial directions, a condition that is satisfied by the shape of the eye, where specific patterns of responses would be generated across the retina (Ritz et al. 2000).

While the eye is clearly indicated as the site of magnetoreception (W. Wiltschko et al. 2002, 2003), and an involvement of a ‘radical-pair’ mechanism is now supported by experimental evidence (Ritz et al. 2004), the wavelength dependency observed in orientation experiments argues against a role for the four color receptors present in cones (W. Wiltschko and R. Wiltschko 2002). Most likely, a different type of photopigment acts as transducer molecule.

Cryptochromes, a class of photoactive pigments absorbing light in the blue to green part of the spectrum have been proposed as candidate molecules for the ‘radical-pair’ model (Ritz et al. 2000). Cryptochromes are flavoproteins, known to be involved in the regulation of circadian rhythms in plants and animals (see Cashmore et al. 1999; Sancar 2003). They possess a number of biochemical properties that are crucial for magnetoreception, including the ability to form radical pairs (Giovani et al. 2003).

In view of this, we searched for cryptochrome transcript expression in the retina of birds belonging to a migrating species. We here report the isolation of three new forms of cryptochrome from a migratory bird, the European robin (Erithacus rubecula).

Materials and methods

Fourteen test birds were mist-netted in the Botanical Garden in Frankfurt am Main (50°08′N, 8°40′E) in September 2001, 2002, and 2003. They were kept in individual cages over the winter in a photoperiod simulating local conditions. In December, the photoperiod was reduced to a light:dark (L:D) cycle of 8:16 h. Starting in January, premature Zugunruhe (migratory restlessness) was induced by increasing the cycle to 13:11 h, with lights turned on at 0400 hours and switched off at 1700 hours Central European Time (CET), since test birds were also used for orientation experiments, in order to verify that they show normal migratory behavior.

Retinal samples were collected on 26 March 2002, 31 March 2003, and 2 March 2004. Birds were decapitated in total darkness, and the retina of each bird was dissected under red light (peak wavelength 720 nm; Philips photo bulb). A piece of the flight muscle (musculus pectoralis) served as control tissue. After dissection, retina and muscle tissue were immediately frozen in liquid nitrogen for subsequent isolation of total RNA using TRIzol Reagent (Gibco BRL/Life Technologies).

Total genomic DNA was isolated from Erithacus flight muscle tissue according to a standard protocol (Ender and Schierwater 2003). Degenerate primers used in the first-round PCR were designed from a highly conserved cryptochrome region found in Mus, Xenopus, Gallus, and Coturnix.

Primer sequences were: f-GACCTSTGGATYAGYTGGGAAGAAGGRATGAAGGT and r-GCATGGTTYACCATTGGTTTGGGATARTTRACTC for CRY1 and f-GGCCMTCRTBAGCCGHATGGA and r-ATCCABCCYTCCTGHCTCAGTTG for CRY2. PCR products of genomic cryptochrome were amplified, cloned (Topo-TA, Invitrogen), and sequenced. Subsequently, Erithacus-specific 5′/3′ RACE primers for cryptochrome 1 and 2 were designed: for CRY1, f-AGAGCTGTTACTTGATGCAGATTGGA and r-TTGCAGCCTTCTGGATGCTCTCTG; for CRY2, f-CCAAAGAAGCCAGTGAGCA.

Total RNA isolated from the robin retina was DNase I (Roche) treated for 15 min at 37°C and used in full-length cDNA synthesis (GeneRacer Kit combined with SuperScript III, Invitrogen). Full-length cDNA sequences of retinal cryptochromes were obtained by means of 5′ and 3′ RACE using Erithacus-specific primers and GeneRacer 5′/3′ primers. RACE fragments were recovered by agarose gel electrophoresis, cloned into the TOPO-plasmid vector (Topo-TA, Invitrogen), and sequenced in both directions in a capillary sequencer (Amersham, MEGABASE 500). The genomic coding regions of eCRY1a and eCRY1b were obtained using a Genome Walker kit (Clontech). PCR products of overlapping walks were sequenced directly.


We isolated three cryptochrome genes, two cryptochrome 1 (CRY1) and one cryptochrome 2 (CRY2) gene, from the retina of robins. Based on the similarity to the known sequences from chicken cCRY1 and gCRY2 (Haque et al. 2002; Bailey et al. 2002), we named them eCRY1a (GenBank accession number AY585716), eCRY1b (accession no AY585717) and eCRY2 (accession no AY772689), respectively (Fig. 1).
Fig. 1

Alignment of C-terminal amino acids of domestic chicken (Gallus gallus) cryptochromes with retinal cryptochrome proteins of the European robin (Erithacus rubecula). Homologue amino acids are indicated by dashes. Upper part shows the alignment of 66 amino acids of Gallus gallus cryptochrome1 (cCRY1) with Erithacus rubecula-specific cryptochromes eCRY1a and eCRY1b. Lower part depicts the alignment of 49 amino acids of Gallus gallus cryptochrome 2 (gCRY2) with Erithacus rubecula cryptochrome 2 (eCRY2). Identified nuclear localization signal (NLS) is shown in bold letters

Characterization of genomic sequences revealed that eCRY1a and eCRY1b originate from the same gene locus and thus are splice products (Fig. 2). As in all other known cryptochromes, the N-terminal regions contain conserved regions of the flavin- and pterin-binding domains (Miyamoto and Sancar 1998; Cashmore et al. 1999; Zhu and Green 2001; Haque et al. 2002; Sancar 2003). The different C-termini of eCRY1a and eCRY1b protein sequences are encoded in different exons (Fig. 2). The first isolated cryptochrome 1, eCRY1a, is homologous to cCRY1 known from domestic chicken (Haque et al. 2002), whereas the other cryptochrome, eCRY1b, differs in having a novel carboxy (C)-terminus domain (Fig. 2).
Fig. 2

Two types of cryptochrome 1, eCRY1a and eCRY1b, are found in the retina of European robins (Erithacus rubecula). They are splice products of the same gene; both possess a conserved N-terminal region, similar to those of other vertebrate cryptochromes (represented by the white bars). The C-terminal regions of both proteins are shown, starting from the last shared amino acids. The eCRY1a-cDNA is homologous to cCRY1-cDNA found in chicken, involving the genomic DNA regions 2027–2042, 3180–3251, and 3350–3450. In contrast, eCRY1b-cDNA codes for a novel C-terminus involving the region 2027–2116 (shaded box)


Our study clearly shows the existence of three forms of cryptochrome expressed in the retina of a migratory passerine bird. For cryptochrome 1, the novel carboxy (C)-terminus domain of eCRY1b suggests different properties than those of eCry1a or other known cryptochromes. Previous studies have shown that the C-terminal regions are especially important for protein-protein interactions (e.g., Rosato et al. 2001).

Although not all cryptochromes were found to be directly activated by light (Griffin et al. 1999; Van Gelder et al. 2003), and not all ‘radical-pair’ processes in flavoproteins have immediate physiological functions, the fact that cryptochromes are found in the robin’s retina is very intriguing in view of the ‘radical-pair’ model (Ritz et al. 2000). Failure to find cryptochromes in the robins’ retina would have made the involvement of this type of photopigment in magnetoreception rather unlikely. Hence, the presence of eCRY1a, eCRY1b, and eCRY2 allows us to speculate that cryptochrome might serve another function besides its well-known role in clock mechanisms, namely that it may be the crucial transducer molecule mediating magnetic compass information. However, we consider an involvement of eCRY2 in magnetoreception less likely because we were able to identify a known nuclear localization signal (PKRKHE) in the C-terminal part of the amino acid sequence (see Fig. 1, lower part). In order to act as a sensory light receptor we would rather expect cryptochromes not to be located in the nucleus.

Cryptochromes have been found in the retina of several vertebrate species (Miyamoto and Sancar 1998; Zhu and Green 2001; Bailey et al. 2002; Haque et al. 2002), where they appear to be evenly expressed over the inner nuclear layer of the retina (Miyamoto and Sancar 1998; Zhu and Green 2001; Haque et al. 2002). They are associated with the so-called displaced ganglion cells (Miyamoto and Sancar 1998; Mouritsen et al. 2004). These cells are distributed across the entire retina (Nalbach et al. 1993), thus fulfilling the requirement of representing various spatial directions, which is crucial for the ‘radical-pair’ model. In addition, axons from these cells project into the nucleus of the basal optic root (nBOR), where electrophysiological responses to changes in magnetic directions have been recorded in pigeons and passerine species (Semm et al. 1984).

These observations, together with theoretical considerations (Ritz et al. 2000) and our findings of CRY in the retina, seem to point to a role of cryptochromes in the primary processes of magnetoreception, although the specific role of the two different types of cryptochrome 1 and the transduction mechanisms remain to be determined. By initiating further studies to unravel details about the specific role of cryptochromes in migratory birds, this study provides a stepping stone in understanding the perception of magnetic compass information.


We gratefully acknowledge laboratory support from Stephen L. Dellaporta and Maria A. Moreno of Yale University during the early months of this study. We thank Allen G. Collins for comments on an earlier draft of this manuscript and Agrani Rump for isolating parts of the eCRY2 sequence. Our work was supported by the Deutsche Forschungsgemeinschaft (grants to W.W. and B.S.), the Human Frontier Science Program (grant to B.S.) and the Studienstiftung des Deutschen Volkes (doctoral fellowship to A.M). The study was performed according to the laws and regulations on animal welfare in Germany.

Copyright information

© Springer-Verlag 2004