A novel isoform of cryptochrome 4 (Cry4b) is expressed in the retina of a night-migratory songbird

The primary sensory molecule underlying light-dependent magnetic compass orientation in migratory birds has still not been identified. The cryptochromes are the only known class of vertebrate proteins which could mediate this mechanism in the avian retina. Cryptochrome 4 of the night-migratory songbird the European robin (Erithacus rubecula; erCry4) has several of the properties needed to be the primary magnetoreceptor in the avian eye. Here, we report on the identification of a novel isoform of erCry4, which we named erCry4b. Cry4b includes an additional exon of 29 amino acids compared to the previously described form of Cry4, now called Cry4a. When comparing the retinal circadian mRNA expression pattern of the already known isoform erCry4a and the novel erCry4b isoform, we find that erCry4a is stably expressed throughout day and night, whereas erCry4b shows a diurnal mRNA oscillation. The differential characteristics of the two erCry4 isoforms regarding their 24-h rhythmicity in mRNA expression leads us to suggest that they might have different functions. Based on the 24-h expression pattern, erCry4a remains the more likely cryptochrome to be involved in radical-pair-based magnetoreception, but at the present time, an involvement of erCry4b cannot be excluded.

Cry4 is suggested to be the most likely candidate for being the magnetoreceptive protein: Not only is Cry4 of night-migratory European robins (Erithacus rubecula) located in a very suitable site for magnetoreception 31 , but it is also upregulated during the migratory season and shows only weak 24-h rhythmicity 50 . Even more importantly, stably bound FAD in vitro has been seen in Cry4 from the chicken 62,63 and the pigeon [64][65][66] and stably bound FAD in vivo is a crucial prerequisite for a protein potentially acting as a magnetoreceptor 27,31 . Furthermore, Cry4 has been suggested to interact with the long-wavelength opsin, iodopsin 67 . Iodopsin is located exclusively in the double cones and long-wavelength single cones, which is where Cry4 is primarily located within the retina of the European robin 50 , and which would seem to be a very suitable site for magnetoreception 31 . Since over 90% of all human protein coding genes produce multiple mRNA isoforms (or splice variants) 68 and two variants of the Cry1 gene, Cry1a and Cry1b, have already been found, we were interested if there might also be more than one isoform of avian Cry4. Here, we report that erCry4 is expressed in at least two splice variants: In addition to the already known erCry4 isoform, now called erCry4a, the novel isoform, called erCry4b, possesses one additional exon located within the DNA photolyase homology region. We found the Cry4b isoform in the retina from several other bird species and characterised the two isoforms concerning their 24-h mRNA expression and tissue distribution in the European robin.

Results
Cry4b is a splice variant of avian Cry4. By analysis of the RT-PCR fragment length and subsequent Sanger sequencing of retinal cDNA of the European robin, we detected a novel isoform of erCry4, which we named erCry4b. In contrast to erCry4 (now named erCry4a), the erCry4b mRNA comprises an additional exon of 87 nucleotides (nt) length (29 amino acids, aa) at nt position 532 to 618, which is characteristic for erCry4b (Fig. 1A, blue). This newly discovered exon is formed by retention of exon 3 of the erCry4 gene in the erCry4b isoform. The open reading frame (ORF) for the erCry4b isoform is composed of eleven exons and has a total length of 1671 nt (556 aa). The mRNA sequence of erCry4b is deposited under the GenBank accession number MN709784. By RT-PCR, we also detected the two isoforms expressed in the retinae from the Eurasian blackcap (Sylvia atricapilla) and the zebra finch (Taeniopygia guttata) (Fig. 1B). In the domestic chicken, we only observed a very faint band of the appropriate size for the Cry4b isoform, which suggests that Cry4b might also be expressed in the chicken, but possibly to a much lower extent (Fig. 1B). When subjecting the novel erCry4bspecific exon to a genomic BLAST database search (https ://blast .ncbi.nlm.nih.gov/Blast .cgi), we received positive hits for a number of other listed bird species (Table 1). This suggests that the Cry4b isoform is most likely conserved among birds.
Tissue distribution and diurnal expression pattern. Using RT-PCR, we monitored the mRNA expression pattern of all five cryptochromes known so far in ten European robin tissues (retina, brain, heart, kidney, liver, lung, spinal cord, muscle, beak, and skin), with the TATA-box binding protein (Tbp) as a reference. All cryptochromes were found to be expressed in all European robin tissues examined, even if the expression of the www.nature.com/scientificreports/ two Cry4 transcripts was extremely low in some of the tissues (heart, kidney, spinal cord and muscle) ( Fig. 2 and Fig. S1). We recently analysed the expression pattern over the course of 24 h in four different European robin retinal cryptochrome transcripts, erCry1a, erCry1b, erCry2 and erCry4, by quantitative RT-PCR 50 . Here, the mRNA of retinal erCry4 (which included both isoforms erCry4a and erCry4b) did not show a pronounced 24-h rhythmicity (Fig. 3, bottom right). Separated examination of the erCry4 isoforms in this study revealed that the transcripts of erCry4a and erCry4b differ in their expression pattern: While erCry4a mRNA did not display a clear and significant 24-h rhythmicity in the retina (Fig. 3, top left; one-way ANOVA: p = 0.152, Cosinor analysis: p = 0.274), we observed a pronounced 24-h rhythm of erCry4b mRNA expression (Fig. 3, top right; one-way ANOVA: p < 0.001), with a peak in the afternoon (13:00 and 16:00 CET) and the lowest values at night (22:00, 1:00 and 4:00 CET). The increase in relative expression is about threefold (Tukey-HSD: p < 0.01) and statistically significant rhythmicity of erCry4b was confirmed by Cosinor analysis (p < 0.001). As expected, adding the values obtained for erCry4a and erCry4b, respectively, results in a graph (Fig. 3, bottom left) that is very similar to the pooled Cry4 mRNA expression pattern from our recent publication 50 (Fig. 3, bottom right). www.nature.com/scientificreports/ The abundance of cryptochrome mRNA was measured by RT-PCR with cDNA from ten different tissues taken from one bird individual, and the −RT control as a negative control (see "Materials and methods"). The TATAbox binding protein (erTbp) served as a reference gene. Full-length gels are provided in the supplement (Fig. S1).

Discussion
Avian Cry4 has been suggested to be the primary magnetoreceptor in the retina of night-migratory birds 31,50 .
Here, we report that the Cry4 gene of the night-migratory European robin is expressed in at least two isoforms, which we then analysed with respect to their distribution in tissues and their 24-h rhythmicity in the retina. The longer isoform, erCry4b, might represent the full-length protein, while the shorter (formerly called erCry4) isoform erCry4a is generated by exon-skipping of exon 3 in the course of an alternative RNA splicing process. The Cry4b-specific exon seems to be conserved among birds since we also detected it in the retinae of the night-migratory Eurasian blackcap, the non-migratory songbird zebra finch and possibly the domestic chicken. Moreover, it is present in the genome of the majority of the bird species contained in BLAST databases. We found a diverging 24-h expression pattern of retinal erCry4a and erCry4b mRNA: While erCry4a seems to be stably expressed and shows no clear rhythmicity in the retina, erCry4b displays a significant 24-h oscillation with its peak in the afternoon (16:00), which corresponds to the expression maximum of both erCry1a and erCry1b mRNA 50 . The differences in the 24-h expression pattern of the two erCry4 isoforms likely explain why we only detected a subtle rhythmicity in our recent study that did not distinguish between erCry4a and erCry4b mRNA 50 . However, we cannot make any predictions about how and when the proteins are functionally active since we only look at the mRNA production level with the qRT-PCR method applied in this study. When comparing the mRNA levels of the two Cry4 isoforms in Figs. 1B, 2 and 3, there is an inconsistency of the relative abundancy of Cry4a and Cry4b: In Figs. 1B and 2, the mRNA levels of Cry4b seem to be higher than those of Cry4a in Fig. 3. This inconsistency is most likely due to the different primer pairs used for RT-PCR (the same oligonucleotides detected Cry4a and Cry4b; Figs. 1B and 2) and qRT-PCR (specific primers were used to detect the isoforms separately; Fig. 3), respectively (see Tables 2 and 3). Moreover, although Cry4b is mainly expressed during the day, the absolute abundance of the Cry4a mRNA at this time point might still be higher.
Cry4 has been proposed to be the putative primary magnetoreceptor underlying the avian magnetic compass 50,66 . In contrast to Cry1 and Cry2, which are most probably unable to bind the cofactor FAD required for the suggested radical-pair mechanism of magnetoreception 62,69 , Cry4(a) from several species has been shown to bind FAD well [62][63][64] . Our present finding of a second Cry4 isoform (here named Cry4b), however, raises the question whether both or only one of them might be involved in magnetoreception. The fact that they display a different 24-h mRNA expression pattern in the retina suggests that erCry4a and erCry4b have different functions within the retina. The lack of 24-h rhythmicity of the erCry4a mRNA is consistent with what would be predicted for a stably expressed retinal magnetoreceptor. On the other hand, erCry4b, displaying a clear diurnal mRNA oscillation, might have a role diverging from Cry4a function: It could either be a component of the circadian clock, share a putative magnetoreceptor function with erCry4a, or play a completely different, so-far unknown role in the retina. Future studies involving functional expression of Cry4a and Cry4b proteins followed by comparative analysis of their photoreactions could potentially answer this question.
We also examined the tissue-wide mRNA distribution of all known cryptochromes in the European robin and detected a widespread expression of all cryptochromes in all tissues examined. This result is consistent with very recent findings of clCry1, clCry2 and clCry4 in eleven tissues of the pigeon 65 . The ubiquitous tissue expression of erCry1a, erCry1b and erCry2 is not in conflict with their probable role in the circadian clock, Table 2. Primer sequences and predicted size of the amplified product for the different genes examined for the tissue-wide distribution study.  www.nature.com/scientificreports/ since both mammalian Cry1 and Cry2 70 as well as other circadian clock proteins have also been observed to be widely expressed throughout mammalian tissues [71][72][73] . A universal tissue distribution of an assumed primary magnetoreceptor, Cry4, on the other side, which is thought to play its key role in the avian retina, seems not very plausible at first sight. However, many proteins perform multiple functions 74 . This phenomenon, termed 'protein moonlighting' or 'gene sharing' , means that one gene can produce proteins with several different molecular functions [74][75][76] . The moonlighting effect may manifest itself through expression in different cell types or different locations within a cell 75 . Not only might one and the same Cry4 isoform play different roles in different tissues, but both Cry4 isoforms could also have different functions (both within the retina and throughout the body).
To find out more about possibly diverging roles of the two Cry4 isoforms in the avian retina, it will be necessary to produce antibodies specifically recognising either Cry4a or Cry4b protein.
In conclusion, we found an additional isoform (Cry4b) of the putative magnetoreceptor Cry4 in the retina of several bird species. The mRNA of this novel isoform is primarily expressed during the day whereas the mRNA of the formerly known Cry4(a) is expressed at similar level throughout a 24-h period. This could indicate that the two isoforms play different roles within both the retina and/or the whole avian organism.

Materials and methods
Birds. Twenty-six European robins (Erithacus rubecula) and two Eurasian blackcaps (Sylvia atricapilla) were wild-caught in the vicinity of the campus of the University of Oldenburg using mist nets. Two zebra finches (Taeniopygia guttata) were captive-bred. Two chickens (Gallus gallus domesticus) were bred from specific-pathogen-free eggs (VALO Biomedia, Osterholz-Scharmbeck, Germany) and raised in the animal care facility of the University of Oldenburg. Of the 26 European robins used in the qRT-PCR study, 24 were the same that were used in ref. 50 (Table S1, bird number 1-24 in ref. 50 ). All animal procedures were performed in accordance with local and national guidelines for the use of animals in research and were approved by the Animal Care and Use Committees of the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit (LAVES, Oldenburg, Germany), Az: 3314-42502-04-10/0121. Tissue collection, RNA isolation and cDNA production. The tissues were collected from birds sacrificed by decapitation. One retina from each bird, free of vitreous, and the other tissues used in this study were put into ice-cold TRIzol Reagent (Life Technologies, Carlsbad, CA, USA), shock-frozen in liquid nitrogen and stored at − 80 °C until RNA extraction. The total RNA was isolated according to manufacturer's instructions. RNA concentration was measured using the Infinite 200 PRO instrument (Tecan, Männedorf, Switzerland). RNA quality was determined with the Agilent RNA 6000 Nano Kit using a 2100 Bioanalyzer Instrument (Agilent Technologies, Santa Clara, CA, USA). For each sample, 1 μg total RNA was reverse-transcribed using the QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany), which included a genomic DNA removal step, according to the manufacturer's instructions. As a negative control, samples from each batch of reverse transcription were pooled and incubated with primer mix lacking reverse transcriptase (− RT control). The obtained cDNA was diluted 1/20 in 0.1 × TE buffer and stored at − 20 °C until use.
Amplification and cloning of the erCry4b cDNA sequence. Specific primers with restriction sites were designed to amplify the coding region of the European robin erCry4 mRNA sequence (gene bank accession number KX890129 50 ). The primers erCry4-Xho1-pT-AE-F, 5-ggtattctcgagatgctgcatcgcaccat-3′ and erCry4-KpnI-pT-AE-R, 5′-tattggtaccgtgtattctgttgttcggg-3′ were used for polymerase chain reaction (PCR) amplification with the Phusion High-Fidelity DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA; Cat. No. F-530XL) following the manufacturer's instructions. The cycle conditions were: one cycle denaturation at 98 °C (30 s), 30 cycles at 98 °C (10 s), 64 °C (30 s), 72 °C (60 s), with a final extension of 10 min at 72 °C. The PCR product was run on a 2% agarose gel for 120 to 150 min at 90 V. The two erCry4 isoforms erCry4a and erCry4b had separated on the gel due to their size difference. The PCR products were cut out from the gel, purified using the QIAquick Gel Extraction Kit (Qiagen) and subjected to sequencing (LGC, Berlin, Germany). Sequence analysis was carried out using Clustal Omega from the European Bioinformatics Institute (https ://www.ebi.ac.uk/Tools /msa/ clust alo/). After detection of a longer isoform (erCry4b), the respective PCR product was digested with the XhoI/KpnI Fast Digest restriction enzymes (Thermo Fisher Scientific) and cloned into the expression vector pTurboGFP-N (Evrogen, Moscow, Russia) according to the protocol for the Rapid DNA Dephos & Ligation Kit (Roche Diagnostics, Rotkreuz, Switzerland). To amplify erCry4b from the retinae of European robins, Eurasian blackcaps, zebra finches and domestic chickens (Fig. 1B), the primers Cry4-F493 (5′-gaggtgcctgtccgaaac-3′) and Cry4-R835 (5′-tcaggcctgtggtgcttg-3′) were used in a PCR reaction with the GoTaq Long PCR Master Mix (Promega), which included pre-incubation at 95 °C for 5 min, followed by 36 cycles of denaturation at 94 °C for 15 s, annealing at 56 °C for 30 s and extension at 72 °C for 30 s, with a final incubation at 72 °C for 10 min. The expected product sizes were 254 bp for Cry4a and 343 bp for Cry4b. Ten µl of a 25 µl PCR amplification assay were subsequently analysed on a 1.7% agarose gel with the BioRad QuickLadder 100 bp as ladder.
RT-PCR tissue distribution. We performed RT-PCR to characterize the gene expression of all known cryptochromes in ten different tissues. The method of RNA extraction and cDNA synthesis was the same as described in the section 'Tissue collection, RNA isolation and cDNA production' . The RNA from all tissues had been subjected to quality analysis using a Bioanalyzer 2000 (Agilent), and only RNA with a RIN > 7.9 was used for cDNA synthesis. Specific primers (listed in Table 2) were designed using Primer 3 (https ://bioin fo.ut.ee/ prime r3-0.4.0/). PCR was performed using GoTaq Long PCR Master Mix (Promega, Madison, WI, USA), which included pre-incubation at 95 °C for 5 min, followed by 32 cycles of denaturation at 94 °C for 15 s, annealing at