Cryptochrome Genes Are Highly Expressed in the Ovary of the African Clawed Frog, Xenopus tropicalis

Cryptochromes (CRYs) are flavoproteins sharing high homology with photolyases. Some of them have function(s) including transcription regulation in the circadian clock oscillation, blue-light photoreception for resetting the clock phase, and light-dependent magnetoreception. Vertebrates retain multiple sets of CRY or CRY-related genes, but their functions are yet unclear especially in the lower vertebrates. Although CRYs and the other circadian clock components have been extensively studied in the higher vertebrates such as mice, only a few model species have been studied in the lower vertebrates. In this study, we identified two CRYs, XtCRY1 and XtCRY2 in Xenopus tropicalis, an excellent experimental model species. Examination of tissue specificity of their mRNA expression by real-time PCR analysis revealed that both the XtCRYs showed extremely high mRNA expression levels in the ovary. The mRNA levels in the ovary were about 28-fold (XtCry1) and 48-fold (XtCry2) higher than levels in the next abundant tissues, the retina and kidney, respectively. For the functional analysis of the XtCRYs, we cloned circadian positive regulator XtCLOCK and XtBMAL1, and found circadian enhancer E-box in the upstream of XtPer1 gene. XtCLOCK and XtBMAL1 exhibited strong transactivation from the XtPer1 E-box element, and both the XtCRYs inhibited the XtCLOCK:XtBMAL1-mediated transactivation, thereby suggesting this element to drive the circadian transcription. These results revealed a conserved main feedback loop in the X. tropicalis circadian clockwork and imply a possible physiological importance of CRYs in the ovarian functions such as synthesis of steroid hormones and/or control of estrus cycles via the transcription regulation.


Introduction
The circadian clock uses properties of both self-sustained oscillation and sensitivity to environmental light for its resetting. In the molecular clock, several clock genes show circadian transcriptional oscillations that are served by positive and negative regulatory factors. CLOCK and BMAL are transcription factors constituting a positive regulatory complex that binds to the CACGTG-type E-box, a core circadian enhancer element [1]. Cryptochromes (CRYs) are unique molecules in that they retain the structure of blue-light photoreceptors, which are highly related to photolyases [2,3], and the vertebrate CRYs operate as negative factors inhibiting E-box-mediated circadian gene transcription [4].
Since the end of the last century, circadian clock components and the circadian clock mechanism have been extensively studied in both vertebrates and invertebrates. However, the studies have mainly concentrated on limited species of mice and fruitflies. Although there are a certain number of reports on clock molecules in zebrafish and Xenopus laevis [5,6], temporal and spatial expression and/or the function of clock molecules in other species, especially in the lower vertebrates, are far less investigated. In this study, we identified two CRYs in the African clawed frog, Xenopus tropicalis, one of the most suitable animal models for genetic manipulations in vertebrates. Further investigation into the expression and circadian function of XtCRYs yielded findings of extremely high expression of cryptochrome mRNA in the ovary of this lower vertebrate.

Results
Based on sequence information from the Ensembl X. tropicalis and EST databases, we cloned full-length cDNAs for XtBmal1, XtCry1, XtCry2 and Xtb2M (b2-microglobulin) from the adult frog kidney. The full-length XtClock sequence has already been identified in the Entrez Nucleotide database (NCBI), and in this study, we were able to determine full-length XtCRY1, XtCry2, XtBmal1 and Xtb2M sequences. The amino acid sequences of the X. tropicalis clock proteins were similar to those of Xenopus laevis, a closely related species of X. tropicalis, indicating that the main frame of circadian clockwork is conserved between the two Xenopus species.
Deduced amino acid sequences of XtCRYs, XtCLOCK, and XtBMAL1 were aligned with their orthologous protein sequences from other species (Figures S1, S2, and S3) and their evolutionary relationships were analyzed using the Neighbor-Joining (NJ) method (CLUSTAL W version 1.83, http://clustalw.ddbj.nig.ac.jp/) (Figures 1, S4, S5 and Table S1). These NJ trees had nearly identical topologies to those constructed using the maximum-likelihood method in the PHYLIP 3.68 software [7] (data not shown).
To uncover whether the mRNA levels of XtCrys varied in a diurnal or nocturnal manner, we performed quantitative RT-PCR using the cDNA collected from various tissues at midday (ZT6) or midnight (ZT18). In order to establish a consistently transcribed gene for reference, we measured the mRNA levels of three genes, b2M, Hprt1 (hypoxanthine-guanine phosphoribosyltransferase1, Genbank accession no. NM_203981) and Gusb (b-glucronidase, Genbank accession no. CT030620). b2M and Hprt1 were selected as the reference control genes for the examination of tissue distribution and diurnal variation, respectively, because the threshold cycle (Ct) values for b2M and Hprt1 remained relatively unchanged for tissues and sampling time, respectively (data not shown). XtCry2 mRNA levels were significantly higher at midday than at midnight in the kidney, muscle, heart, liver, and fat-pad tissues ( Figure 2A). Similar variation was seen in other tissues such as the skin, retina, and stomach. We detected significant variation in XtCry1 mRNA levels in the spleen, retina, stomach, and fat pads. These changes of Cry mRNA expression levels ( Figure 2A) are consistent with a possible circadian function of XtCrys, but these results do not necessarily indicate that there are endogenous XtCry mRNA rhythms, because daily changes in mRNA may be simply driven by light directly or indirectly. Another control gene, XtGusb, showed possible nocturnal change in the ovary under these conditions, although this finding was not statistically significant. When the mRNA levels of the two time points (ZT6 and ZT18) were averaged to compare mRNA expression levels among the tissues, the XtCry mRNA levels were by far the highest in the ovary versus the other tissues examined ( Figure 2B, C). Levels in the ovary were about 28-fold (XtCry1) and 48-fold (XtCry2) higher than levels in the next abundant tissues, the retina and kidney, respectively ( Figure 2B, C).
Next, we performed transcriptional analysis to evaluate whether the putative X. tropicalis clock proteins could constitute a circadian molecular loop. At first, the CACGTG-type E-box element and its related sequences in the promoter region of XtPer1 were examined as a plausible core negative regulator in the circadian molecular loop [8]. Since a CACGTG sequence was found in the 22217,22212 upstream region of XtPer1 ( Figure 3A), a tandem repeat of the 20 bp sequence containing the E-box sequence (corresponding to 22224,22205 region) was used in the luciferase reporter analysis. Coexpression of XtCLOCK and XtBMAL1 showed more than an 81-fold transactivation from the E-box-containing sequence, and the transactivation was strongly inhibited by XtCRY1 or XtCRY2 in a dose-dependent manner ( Figure 3B). Based on reports that nuclear translocation is essential for the E-box-mediated transcription inhibition of CRYs [9], we  Table S1.) were analyzed in the conserved region of the CRY family proteins (corresponding to Arg 10 -Pro 447 in XtCRY1) using the Neighbor-Joining method and CLUSTAL W. CRY DASH proteins (XlCRY DASH, zCRY DASH, and AtCRY DASH) were used as the outgroup (not shown). Bootstrap probabilities (p) are represented by closed circles on the nodes (p.98%) or values near the nodes. Abbreviations are defined as follows: h, human; m, mouse; c, chicken; Xl, Xenopus laevis; Xt, Xenopus tropicalis; z, zebrafish; Ag, Anopheles gambiae; Am, Apis mellifera; Ap, Antheraea pernyi; Dm, Drosophila melanogaster; Dp, Danaus plexippus; PHR, photolyase. doi:10.1371/journal.pone.0009273.g001 investigated possible nuclear localization of the XtCRYs by expressing their GFP-fusion proteins in HEK 293 cells. Both GFP-CRY1 and GFP-CRY2 localized in the nucleus, while GFP alone localized predominantly in the cytoplasm ( Figure 4).

Discussion
In this study, we identified two cryptochrome genes in X. tropicalis and investigated their mRNA expression in various tissues ( Figure 2). Based on the daily variations of both Cry mRNAs found in several tissues (Figure 2A), together with their potential function in circadian transcriptional regulation ( Figure 3), both CRYs may play a key role in the X. tropicalis circadian clock. A comparison of deduced amino acid sequences of the CRYs in the two Xenopus species (Tables S2 and S3) indicated that CRY1 is more highly conserved in the two species (97.4%) than CRY2 (92.8%), implying a slower evolution of Cry1 genes than Cry2 genes. This implication is also supported by a phylogenetic tree of CRY/Photolyase family proteins ( Figure 1) in which vertebrate CRY1 proteins (frog, chicken, and mammalian CRY1) were more closely clustered in comparison with CRY2 proteins. There is only one Cry2 gene in X. tropicalis, whereas there are two Cry2 genes in X. laevis. In the NJ tree (Figure 1), XtCry2 directly clustered with XlCry2b, and XlCry2a positioned itself outside these three Xenopus Cry2 genes. Therefore, it is likely that duplication of the ancestral Cry gene occurred before divergence of the two Xenopus species, although we did not find any Cry2a-like sequences in the X. tropicalis genomic or EST sequences. Alternatively, duplication of the ancestral XlCry gene occurred with a genome duplication of ancestral X. laevis. This speculation is apparently inconsistent with the local tree topology but agrees well with the fact that extant X. tropicalis and X. laevis are diploid and pseudoquadpolyploid, respectively. Further analyses of the genome sequences of both Xenopus species, including the genomic location, structure, and organization of Cry genes, which are yet unavailable in public databases, can help to expand upon these speculations.
In this study, both XtCry1 and XtCry2 transcripts were abundantly expressed in the X. tropicalis ovary ( Figure 2B, C). There are currently few studies on circadian clock gene expression in the ovary of vertebrates, except for mice and zebrafish (see below), and to the best of our knowledge, the present result is the first report on extremely high accumulation of Cry mRNA in the ovary of vertebrates including mice and zebrafish. In mice, mClock, mBmal1, and mCry1 mRNA have been transcribed in the oocyte before fertilization [10,11], and mCry1 transcript level was found to be reduced to very low levels until the end of the 1-cell stage after fertilization, while both mClock and mBmal1 transcript levels were not reduced until the end of the 2-cell stage [11]. Presumably, the reduction of mCry1 transcripts resulted in the relative elevation of mClock and mBmal1 transcripts, and hence transcriptional activation by the CLOCK:BMAL heterodimer would trigger circadian clockdependent regulation of the cell cycle [12] and/or synchronization of the intrinsic developmental clock of mouse oocytes. Similarly, large amounts of XtCry1 and XtCry2 transcripts accumulating in the Xenopus ovary might arrest the cell cycle by inhibiting the E-boxcontrolled cell cycle gene(s). The function of CRY in the ovary is still unknown and will be interesting to be pursued in the future.
In zebrafish, oscillation of Per3 mRNA in the unfertilized oocyte is reported under light-dark (14:10 h) conditions [13], while the expressions of other clock genes are not reported. This zPer3 oscillation may indicate that unfertilized zebrafish oocytes have either the presence of an oscillating circadian clock or the capability of using photoreception as a transcriptional control. Mammalian CRY proteins operate as non-photoreceptive transcriptional repressors, whereas avian and invertebrate CRY are likely to retain both photosensitivity and transcriptional regulatory functions [14][15][16]. Although the present study verified only the transcriptional regulatory function of XtCRY1 and XtCRY2 (Figure 3), it might be possible to hypothesize that these CRYs might serve as photoreceptors in the oocyte and control the cirdcadian clock, cell cycle, or hormone secretion according to environmental light conditions.
Recently, it has been suggested that CRYs operate as magnetoreceptors and use light energy for their molecular function [17]. In fact, dCRY, the blue-light circadian photoreceptor in Drosophila, is required for light-dependent magnetic response [18]. Because magnetoreception has also been reported in the larvae of X. laevis [19], it is important to further characterize XtCRYs by evaluating not only their photoreceptive function but also their possible magnetoreceptive function.
Although amphibian clock genes have been studied in X. laevis [6,8,20], the present identification and initial characterization of clock genes, including Crys in X. tropicalis, is of relative significance. In spite of the close kinship between X. tropicalis and X. laevis, the following make X. tropicalis an advantageous experimental animal model: (i) X. tropicalis is a diploid and the genome size is 1.7 Gb, which is smaller than an allotetraploid X. laevis (3.1 Gb), (ii) the generation time of X. tropicalis (3-4 months) is shorter than that of X. laevis (8-12 months) [21,22], and (iii) X. tropicalis is suitable for large-scale breeding because of its smaller body size compared to X. laevis. These advantages will encourage further analyses of CRYs that will in turn contribute to a wide variety of research in fields such as reproductive biology, circadian biology, photobiology and magnetochemical biology.

Animals
Animals were treated in accordance with the guidelines of WASEDA university. Adult X. tropicalis were entrained in 12 hr light/dark (LD 12:12 h) cycles for at least two weeks. Tissues were collected at Zeitgeber time 6 (ZT6) under fluorescent light (,300 mW/cm 2 ) or at ZT18 under dim red lighting (.640 nm; ,120 mW/cm 2 ), respectively, and kept in RNAlater (Ambion) at 4uC until RNA extraction. X. tropicalis were kindly provided by Dr. Takase and Dr. Yaoita (the National Bio-Resource Project (NBRP) of the MEXT, Japan; Institute for Amphibian Biology, Graduate school of Science, Hiroshima University), or bred from these frogs.

Quantitative RT-PCR Analysis
Total RNA was extracted from the tissues using TRIzol reagent (Invitrogen). Quantitative RT-PCR analyses of 1 mg of total RNA were performed using StepOnePlus (Applied Biosystems) along with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). The primers for quantitative RT-PCR are described in Table 1.

Statistical Analysis
Data were analyzed using the Student's t-test or ANOVA with Tukey-Kramer multiple comparisons in the Statcel2 (the add-in forms on Excel (Microsoft)) software.

Cloning of cDNAs of Clock Genes Encoding Full-Length Coding Sequences
Total RNA was extracted from the kidney using TRIzol reagent (Invitrogen). Equal amounts of total RNA from the ZT6 and ZT18 samples were mixed and first-strand cDNAs were synthesized with SuperScript III reverse transcriptase (Invitrogen) using oligo(dT) 21 primer. Primers for cDNA cloning were designed based on database-deposited partial sequences, which anneal to the untranslated region of XtClock, XtBmal1, XtCry1, XtCry2 and Xtb2M genes (Table 1) or inside the ORF region of XtHprt1 and XrGusb genes. PCR products amplified with PfuUltra (Stratagene) were inserted into the pENTR/D-TOPO vector (Invitrogen), and the inserts of at least three independent clones for each clock gene were sequenced. Full-length cDNAs without presumed PCR errors were isolated.

Transcriptional Assay
HEK 293 cells (RIKEN CELL BANK) were plated 3.0610 5 cells per well on 6-well plates. Plasmid constructs were mixed and transfected into HEK 293 cells using Plus Reagent (Invitrogen) and Lipofectamine Reagent (Invitrogen) according to the manufacturer's instructions. Twenty-four hours after transfection, cell lysates were extracted and transcriptional assays were performed using the Dual-Luciferase assay Kit (Promega).

Localization of CRY Proteins in HEK 293 Cells
Expression vectors for N-terminally GFP-tagged CRY proteins were constructed by transferring Cry cDNA into the pcDNA-DEST53 vector (Invitrogen) using the Gateway system (Invitrogen). The expression vector for GFP alone was constructed by the recombination of pENTR/D-TOPO-empty with pcDNA-DEST53.
HEK 293 cells were plated onto CHAMBER SLIDE (IWAKI). The expression plasmids were transfected into HEK 293 cells using Plus Reagent (Invitrogen) and Lipofectamine Reagent (Invitrogen, described above). Twenty-four hours after transfec-