Unique clockwork in photoreceptor of rat

Authors

  • Katja Schneider,

    1. Institute of Functional and Clinical Anatomy, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany
    Search for more papers by this author
    • Katja Schneider and Susanne Tippmann share first authorship.

  • Susanne Tippmann,

    1. Institute of Functional and Clinical Anatomy, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany
    Search for more papers by this author
    • Katja Schneider and Susanne Tippmann share first authorship.

  • Isabella Spiwoks-Becker,

    1. Institute of Microanatomy and Neurobiology, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany
    Search for more papers by this author
  • Heike Holthues,

    1. Faculty of Computer Science and Engineering, Frankfurt University of Applied Sciences, Frankfurt am Main, Germany
    Search for more papers by this author
  • Tanja Wolloscheck,

    1. Institute of Functional and Clinical Anatomy, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany
    Search for more papers by this author
  • Gabriele Spatkowski,

    1. Institute of Functional and Clinical Anatomy, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany
    Search for more papers by this author
  • Lydia Engel,

    1. Institute of Microanatomy and Neurobiology, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany
    Search for more papers by this author
  • Ute Frederiksen,

    1. Institute of Functional and Clinical Anatomy, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany
    Search for more papers by this author
  • Rainer Spessert

    1. Institute of Functional and Clinical Anatomy, University Medical Center of the Johannes Gutenberg University Mainz, Mainz, Germany
    Search for more papers by this author

Address correspondence and reprint requests to Rainer Spessert, Department of Functional and Clinical Anatomy, University Medical Center of the Johannes Gutenberg University Mainz, Saarstraße 19-21, 55099 Mainz, Germany. E-mail: spessert@uni-mainz.de

Abstract

J. Neurochem. (2010) 115, 585–594.

Abstract

In mammals, the retina contains a clock system that oscillates independently of the master clock in the suprachiasmatic nucleus and allows the retina to anticipate and to adapt to the sustained daily changes in ambient illumination. Using a combination of laser capture micro-dissection and quantitative PCR in the present study, the clockwork of mammalian photoreceptors has been recorded. The transcript amounts of the core clock genes Clock, Bmal1, Period1 (Per1), Per3, Cryptochrome2, and Casein kinase Iε in photoreceptors of rat retina have been found to undergo daily changes. Clock and Bmal1 peak with Per1 and Per3 around dark onset, whereas Casein kinase Iε and Cryptochrome2 peak at night. As shown for Clock, Per1, and Casein kinase Iε, the oscillation of transcript amounts results in daily changes of the protein products. The in-phase oscillation of Clock/Bmal1 with Pers and the rhythmic expression of Casein kinase Iε do not occur in molecular clocks of other tissues including the suprachiasmatic nucleus. Therefore, the findings presented suggest that the photoreceptor clock is unique not only in its position outside the clock hierarchy mastered by the suprachiasmatic nucleus, but also with regard to the intrinsic rhythmic properties of its molecular components.

Abbreviations used:
CKI

Casein kinase I

CkIδ

Casein kinase Iδ

CkIε

Casein kinase Iε

Cry

cryptochrome

DD

constant darkness

HOPE

Hepes-glutamic acid buffer mediated organic solvent protection effect

LD

light/dark

LMPC

laser microdissection and pressure catapulting

Per

Period

SCN

suprachiasmatic nucleus

ZT

Zeitgeber time

Circadian clocks drive daily rhythms in physiology and behavior. They consist of interlocking positive and negative transcriptional/translational feedback mechanisms. In mammals, the clock genes Clock, Bmal1, Period 1–3 (Per1–3), and Cryptochrome 1–2 (Cry1,2) compose the core feedback loop in which a CLOCK : BMAL1 heterodimer binds to a CACGTG E-box and activates transcription of Per(s) and Cry(s) (for reviews, see Ko and Takahashi 2006; Hastings et al. 2008a). Unlike the said clock genes, Casein kinase Iε (CkIε) and Casein kinase Iδ (CkIδ) are reported to be constitutively expressed without diurnal rhythmicity (Takano et al. 2000; Ishida et al. 2001; Lincoln et al. 2002). Casein kinases influence the levels and nuclear translocation of clock gene products by post-translational modification. As a consequence, mutations in CkI dramatically shorten the clock cycle in mice and underlie the familial advanced sleep phase syndrome in humans (Xu et al. 2005, 2007).

Circadian clocks are present in a variety of tissues and cells and are organized in a hierarchical manner. The central clock in the suprachiasmatic nucleus (SCN) lies at the top of the hierarchy and is reset mainly by external light signals captured by the retina (for review, see Hastings et al. 2008a). It synchronizes clocks that occur throughout the body and that are referred to as peripheral clocks. As a unique photosensitive system, the retina possesses its own SCN-independent clock system, which allows the retina to anticipate and adapt to the more than one-million-fold change in light intensity in the environment during a 24-h period (Storch et al. 2007; Ribelayga et al. 2008; for review, see Tosini et al. 2008). In mammals, the retinal clock system consists of clocks localized in multiple types of neurons including photoreceptor cells, horizontal cells, bipolar cells, and ganglion cells (for review, see Tosini et al. 2008). The clock within photoreceptors is of particular interest, because it drives a circadian rhythm in melatonin formation (Tosini and Menaker 1996, 1998; Tosini et al. 1998), supposedly via the expressional control of the penultimate enzyme in melatonin formation, N-acetyltransferase [arylalkylamine N-acetyltransferase (AA-NAT); EC 2.3.1.87] (Sakamoto and Ishida 1998; Sakamoto et al. 2004; for review, see Iuvone et al. 2005). The role of the photoreceptor clock in melatonin formation is of high functional and pathological significance not only for the photoreceptor itself, but also for the retina as a whole. This is evident from observations that melatonin influences multiple physiological and pathological parameters of the vertebrate retina, such as outer segment membrane turnover, neurotransmitter release and light-induced damage sensitivity of photoreceptors and other neurons in the retina (for review, see Iuvone et al. 2005).

To gain insight into a clockwork mechanism located outside the SCN clock system and to reveal the molecular basis of daily adaptation in retinal function, we have, in the present study, investigated the clockwork of the photoreceptor in the intact retina. The findings of this study suggest that the photoreceptors possess a clock that is unique with regard to the intrinsic rhythmic properties of its molecular components.

Materials and methods

Animals

Animal experimentation was carried out in accordance with the European Communities Council Directive (86/609/EEC). Adult male and female Sprague–Dawley rats (body weight: 150–180 g) were kept under standard laboratory conditions (illumination with fluorescent strip lights, 200 lux at cage level during the day and dim red light during the night; 20 ± 1°C; water and food ad libitum) under light/dark 12 : 12 (LD 12 : 12) for 3 weeks. When indicated, the rats were then kept for two cycles under dim red light and killed during the next cycle. Animals were killed at the indicated time points by decapitation following anesthesia with carbon dioxide. All dissections during the dark phase were carried out under dim red light. Retinas were quickly removed and immediately processed as follows.

Sample preparation

The HOPE technique (HOPE, Hepes-glutamic acid buffer mediated organic solvent protection effect; DCS, Hamburg, Germany) was applied for the fixation of the retinas. Briefly, fixation started with the incubation of fresh retinas in an aqueous protection-solution HOPE I (DCS) for 48 h at 0–4°C. Retinas were then dehydrated in a mixture of HOPE II solution (DCS) and acetone for 2 h at 0–4°C, followed by dehydration in pure acetone for 2 h at 0–4° C (repeated twice). Tissues were then embedded with low-melting paraffin (Tm = 52–54°C). Tissue sections (10 μm) from HOPE-fixed and paraffin-embedded retinas were prepared on membrane-mounted slides (DNase/RNase free PALM MembraneSlides, P.A.L.M., Bernried, Germany). Three sections were placed onto each slide. The sections were deparaffinized with isopropanol (2 × 10 min each, at 60°C). All sections were stained with cresyl violet (1% w/v cresyl violet acetate in 100% ethanol) for 1 min at 21°C, washed briefly in 70% and 100% ethanol, and then air-dried (Goldmann et al. 2006).

Laser microdissection and pressure catapulting

In order to isolate photoreceptor cells from the stained sections in a contact- and contamination-free manner, the laser microdissection and pressure catapulting (LMPC) technique was applied. LMPC was performed by using a PALM MicroBeam system (Zeiss MicroImaging, Munich, Germany) with PALM RoboSoftware (P.A.L.M., Bernried, Germany). Under the 10× objective, photoreceptor cells were selected, cut, and catapulted into the caps of 0.5 mL microfuge tubes with an adhesive filling (PALM AdhesiveCaps, P.A.L.M.) by utilizing a pulsed UV-A nitrogen laser. Smaller areas of the sections were pooled to reach total average sample sizes of 4 million square microns per tube. Alternatively, the whole retina was excised with a scalpel and collected in a 0.5 mL microfuge tube. Cell lysis for RNA preparation was carried out immediately after sample collection.

RNA extraction

RNA of the laser-microdissected tissue samples was isolated using the RNeasy Micro Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. Briefly, collected cells were lysed in a guanidine-isothiocyanate-containing buffer (RLT buffer) supplied by the manufacturer. The lysates were diluted with RNase-free water and treated with proteinase K. The samples were then cleared by centrifugation, diluted with ethanol, and applied to an RNeasy MinElute Spin Column in order to bind RNA to the silica-gel membrane. After the first washing step, an on-column DNase treatment with RNase-free DNase I was carried out as described by the manufacturer. Isolated RNA was eluted in a final volume of 12 μL RNase-free water. The amount of extracted RNA was determined by measuring the optical density at 260 nm and 280 nm.

RT and quantitative PCR

cDNA was synthesized by using the Verso cDNA Kit (Abgene, Hamburg, Germany), following the manufacturer’s instructions. Briefly, 4.5 μL RNA solution was reverse transcribed by using anchored oligo-dT primers supplied with the kit in a final volume of 20 μL. cDNA was then diluted 1 : 3 in RNase-free water, and aliquots of 5 μL were used for PCR. Quantitative PCR was carried out in a total volume of 25 μL containing 12.5 μL ABsolute™ QPCR SYBR® Green Fluorescein Mix (Abgene), 0.75 μL of each primer (10 μM), 6 μL RNase-free water, and 5 μL sample. Primer sequences are listed in Table 1. PCR amplification and quantification were performed in an i-Cycler (Bio-Rad, Munich, Germany) according to the following protocol: denaturation for 3 min at 95°C, followed by 40 cycles of 30 s at 95°C, 20 s at 60°C, and 20 s at 72°C. All amplifications were carried out in duplicate. By using agarose gel electrophoresis, the generated amplicons for all genes under examination were shown to possess the predicted sizes (Table 1). The amount of RNA was calculated from the measured threshold cycles (Ct) by a standard curve. The transcript amount of Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was constitutively expressed over the 24-h period in photoreceptors and in preparations of the whole retina. Values were then normalized with respect to the amount of glyceraldehyde-3-phosphate dehydrogenase mRNA present.

Table 1.   Primer sequences
GeneAccession numberPrimer sequence 5′ to 3′Length of PCR product (bp)
GapdhNM017008Forward: ATGACTCTACCCACGGCAAG89
Reverse: CTGGAAGATGGTGATGGGTT
rhodopsinNM033441Forward: ACTCAGAAGGCAGAGAAGG638
Reverse: CCTAGACACTTAGCCGTAAGC
ThNM012740.3Forward: CCAGGGCCTTTCCCAAAGTC186
Reverse: CATGGAGGGCAGGAGGAATG
Bmal1AF317669Forward: TCCACAGCACAGGCTACTTGAA105
Reverse: TTGCAACGAGGCAGCTCAGAT
ClockNM021856Forward: GACAGCCCCACTGTACAATACG81
Reverse: TGCGGCATACTGGATGGAAT
Per1XM340822Forward: TGGTAAAGCACCAGGGACAAC102
Reverse: GAAGGACTTTGGCCTTGAATGTAC
Per2NM031678Forward: AGCAGTCCCCTACAGCTTAACCT129
Reverse: CCGAGATGCGCCAGATGT
Per3NM023978Forward: CCGGGATGTGTGTTTCTTGAAG133
Reverse: TGGTGTACTGCAACCATCAGAG
Cry1AF545855Forward: GGGAAGCGCCCAAGTCAGGAAG91
Reverse: CCTCCCGCATGCTTTCATATC
Cry2NMI33405Forward: TTCAGAAGGCCGCTAATTGC102
Reverse: AGATCTGCTTCATCCGCTCAA
CkIεNM_031617Forward: GCAAGCTGGCAATACTTCTC165
Reverse: TCAAATGGCACGCTTGTC
Aa-natU38306Forward: GAAGGGAGACAGCAGTTC431
Reverse: GTCCTGGTCTTGCCTTTTG

Immunoprecipitation and western blot

For immunoprecipitation of Casein kinase Iε (CK1ε), two retinas were homogenized in 1 mL phosphate buffered saline containing protease inhibitors. Insoluble material was pelleted. For antibody immobilization, protein A–agarose beads (30 μL bead volume; Invitrogen, Carlsbad, CA, USA) were washed four times with mouse anti-CK1ε monoclonal antibody (diluted 1 : 100; Santa Cruz Biotechnology, Heidelberg, Germany) at 4°C. Cell extracts corresponding to 250 μg protein amount were applied overnight to the antibody-coupled beads at 4°C. Bound proteins were recovered after extensive washes in homogenization buffer. For western blot analysis, samples were loaded on 4–12% NuPAGE Novex Bis-Tris gels (Invitrogen), separated and blotted onto nitrocellulose transfer membranes (Protran, Schleicher & Schüll, Dassel, Germany). For immunodetection, membranes were blocked in 5% skimmed milk powder and the mouse anti-CK1ε monoclonal antibody (1 : 200) was applied overnight at 4°C. The horseradish-peroxidase-coupled secondary antibodies (goat anti-mouse-horseradish peroxidase 1 : 10 000; Dianova, Hamburg, Germany) were visualized using an ECL detection system (GE Healthcare Amersham, Freiburg, Germany).

Light microscopic immunocytochemistry

Immunostaining of the photoreceptors was investigated at the indicated time-points. For each time point, two animals from three or more independent experiments were analyzed. Labeling was performed by the indirect fluorescence method (for details, see Spiwoks-Becker et al. 2008). The following primary antibodies were used: rabbit anti-CKIε (diluted 1 : 200; Santa Cruz Biotechnology) and mouse anti-CtBP2 (1 : 1000; BD Biosciences, Heidelberg, Germany). The binding sites of the primary antisera were revealed by secondary antisera: goat anti-rabbit (Alexa 594; 1 : 500) or goat anti-mouse (Alexa 488; 1 : 500; both from Molecular Probes, Eugene, OR, USA). At least three randomly chosen labeled sections were examined per animal and retina and photographed using a confocal laser scanning microscope (LSM SP5, Leica Microsystems, Heidelberg, Germany). Multiple pairs of optical sections were recorded at different depths throughout the tissue sections and merged into a single plane (Image-Pro Plus 5.0, Media Cybernetics, Silver Spring, MD, USA). Photoshop software (Adobe Photoshop 7.0, San Jose, CA, USA) was used to edit single pictures and to assemble figures.

Statistical analysis

All data are given as the mean with standard error of the mean (SEM) of five (PCR) or three (western blot) independent experiments. Non-parametric analysis of variance (Kruskall–Wallis test) and, when appropriate, one-way anova was used to reveal differences in the 24-h profile of each transcript. Cosinor analysis (Lentz 1990) was used to determine the acrophase (peak expression) and amplitude of oscillation (half the difference between the highest and lowest values). Student’s t-test was performed to determine differences between the two groups in western blot analysis. All statistical analyses should be regarded as being explorative and p-values are given descriptively with no significance level being fixed. Statistical analysis was performed using SigmaStat (Version 3.10, Systat Software Inc., San Jose, CA, USA), with the exception of Cosinor analysis that was performed using the statistic software ‘R’ (version 2.11.1, freely available at http://www.r-project.org).

Results

Isolation of photoreceptors

To obtain their transcriptomes, photoreceptors were collected from rat retina using the LMPC technique (Fig. 1). To verify the purity of the preparations, photoreceptors were subjected to molecular analysis with rhodopsin as a marker for photoreceptors and tyrosine hydroxylase (Th) as a marker for inner retinal neurons. In photoreceptors collected by LMPC, the ratio of rhodopsin to tyrosine hydroxylase was increased 55 (± 12)-fold compared with preparations of the whole retina.

Figure 1.

 Photomicrographs of a retinal section before (a) and after (b) the photoreceptors were captured by the laser microdissection and pressure catapulting (LMPC) technique. OS, outer segments of photoreceptors; IS, inner segments of photoreceptors; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer.

24-h profiles in mRNA amount of Aa-nat in photoreceptors and preparations of whole retina

To test the functioning of the photoreceptor clock and to adjust our experimental system to previous studies, the 24-h regulation of Aa-nat expression (Fig. 2) was investigated in the same transcriptomes as those used for monitoring the expression of clock genes (Fig. 3). In photoreceptors, the amount of Aa-nat transcript displayed a marked daily rhythm (= 0.005) with an amplitude of 39.33% and acrophase (peak expression) at Zeitgeber Time 17.20 (ZT17.20, i.e., 17.20 h after ‘lights on’) (Fig. 2). Thus, transcription of Aa-nat was ∼ 30-fold higher in the middle of the dark phase than in the middle of the light phase. The daily rhythm in the amount of Aa-nat transcript was retained after three cycles of DD (constant darkness) in photoreceptors (p = 0.012; amplitude: 34.11%; acrophase: ZT0.52), although the rhythm in the amount of Aa-nat mRNA was phase-delayed in comparison with LD.

Figure 2.

 Daily profiles in expression of amount of Aa-nat mRNA under light/dark (LD) 12 : 12 (left column) and constant darkness (DD) (right column). The white circles/solid lines represent photoreceptors (PRC), and the black circles/dashed lines represent preparations of the whole retina. The amount of AA-NAT transcript displays 24-h changes in photoreceptors and in preparations of whole retina under both LD and DD conditions (ZT, Zeitgeber time; CT, circadian time). The solid bars indicate the dark period. Data represent a percentage of the maximal value of transcript amount in photoreceptors and preparations of whole retina. Each value represents mean ± SEM (n = 5, each n from six pooled retinae).

Figure 3.

 Daily profiles in expression of clock genes under light/dark (LD) 12 : 12 (left column) and constant darkness (DD) (right column). The white circles/solid lines represent photoreceptors (PRC), and the black circles/dashed lines represent preparations of whole retina. Under LD conditions, 24-h oscillations were observed in the amounts of RNA for Clock, Per1, Per3, Cry2, and CkIε in photoreceptors and for Clock, Bmal1, Per1, Per2, Per3, and CkIε in preparations of whole retina. Under DD, 24-h oscillations were observed in the transcript amounts of Clock, Per3, and CkIε in photoreceptors and of Clock, Bmal1, Per2, Per3, and CkIε in preparations of whole retina. The solid bars indicate the dark period (ZT, Zeitgeber time; CT, circadian time). Data represent a percentage of the maximal value of transcript amounts in photoreceptors and preparations of whole retina. Each value represents mean ± SEM (n = 5; each n from six pooled retinae).

As in photoreceptors, the amount of Aa-nat transcript in preparations of the whole retina also exhibited daily variations (p = 0.004; amplitude: 21.61%). The transcript amount of Aa-nat peaked at ZT17.08 with a day/night ratio of ∼ 25. After three cycles of DD, the daily variation in the amount of Aa-nat transcript tended to persist (p = 0.076; amplitude: 12.34%; acrophase: ZT17.38), although, in comparison to light/dark conditions, the day/night ratio was clearly decreased.

24-h profiles in transcript amounts of clock genes in photoreceptors

To test and specify the functioning of the photoreceptor clock, the expression of the clock genes Clock, Bmal1, Per1, Per2, Per3, Cry1, Cry2, and CkIε in photoreceptors under LD 12 : 12 was investigated (Fig. 3, left column). All clock genes examined were expressed in photoreceptors. Daily variations were recorded in the amount of mRNA of Clock (p = 0.0035; amplitude: 26.67%; acrophase: ZT11.28), Bmal1 (p = 0.097; amplitude: 23.27%; acrophase: ZT12.40), Per1 (p = 0.028; amplitude: 38.42%; acrophase: ZT11.11), Per3 (p = 0.002; amplitude: 47.70%; acrophase: ZT11.22), Cry2 (p < 0.024; amplitude: 17.66%; acrophase: ZT19.22), and CkIε (p = 0.023; amplitude: 26.18%; acrophase: ZT21.04). The transcript amount of the clock genes under daily regulation increased gradually during the light phase and peaked around dark onset with the exception of Cry2 and CkIε, which showed maximum expression during the dark phase (Fig. 3, left column). Among all clock genes, Per3 displayed the most prominent daily change in expression. The amount of Per3 transcript increased ∼ 20-fold at dark onset in comparison with light onset. After three cycles of DD, daily variations in the transcript amount were retained only for Clock (p = 0.019; amplitude: 26.84%; acrophase: ZT7.52) and Per3 (p = 0.009; amplitude: 36.56%; acrophase: ZT9.47) (Fig. 3, right column).

24-h profiles of transcript amount of clock genes in preparations of whole retina

To allow a comparison between the retina as a whole and photoreceptors and also to normalize the experimental system used with respect to previous studies, the amount of transcript of the above-mentioned clock genes was monitored in preparations of whole retina under LD 12 : 12 (Fig. 3, left column). All clock genes under investigation were expressed in the retina, with 24-h changes in the transcript amount being observed for Clock (p = 0.041; amplitude: 26.80%; acrophase: ZT10.14), Bmal1 (p = 0.048; amplitude: 25.71%; acrophase: ZT9.34), Per1 (p = 0.03; amplitude: 39.11%; acrophase: ZT11.26), Per2 (p = 0.006; amplitude: 22.73%; acrophase: ZT14.04), Per3 (p = 0.005; amplitude: 40.07%; acrophase: ZT10.23), and CkIε (p = 0.024; amplitude: 11.95%; acrophase: ZT17.54). In general, the expression of all clock genes under daily regulation increased gradually during the light phase and peaked around dark onset. The only exception was CkIε, which peaked at night. After three cycles of DD, daily changes in expression were maintained for Clock (p = 0.049; amplitude: 30.86%; acrophase: ZT10.05), Bmal1 (p = 0.033; amplitude: 26.04%; acrophase: ZT10.28), Per2 (p = 0.003; amplitude: 30.04%; acrophase: ZT12.05), and Per3 (p < 0.001; amplitude: 38.21%; acrophase: ZT10.35) (Fig. 3, right column).

Daily changes in CKIε protein amount

To investigate whether daily regulation of CKIε is also evident at the protein level, the amount of CKIε protein was compared for ZT6, ZT18, and ZT0 using western blot analysis (Fig. 4a and b) and laser scanning microscopy (Fig. 4c). Western blot analysis of preparations of the whole retina revealed a protein band of ∼ 42 kDa characteristic for CKIε (Fig. 4a). The intensity of CKIε immunoreactivity underwent daily changes with stronger immunoreactivities at ZT18 (p = 0.002) and ZT0 (p = 0.057) when compared with ZT6 (Fig. 4b). Laser scanning microscopy demonstrated that CKIε was localized primarily in the cytosol of photoreceptors and neurons of the inner retina (Fig. 4c). The intensity of staining displayed daily changes in both types of neurons that paralleled those observed in western blotting. These findings indicated that, in whole retina and photoreceptors, the protein amount of CkIε is under daily regulation.

Figure 4.

 Diurnal changes in the immunoreactivity of CKIε in rat retina. (a,b) Western blot analysis of CKIε protein at ZT6, ZT18 and ZT0 in rat retina. Each lane was loaded with CKIε immunoprecipitated from 250 μg homogenate obtained from the pooling of two retinas. Note that the intensity of CKIε immunoreactivity is stronger at ZT18 and ZT0 than at ZT6. (a) Representative western blot showing that the anti-PDE10A antibody recognizes a band of ∼ 42 kDa characteristic for CKIε protein. (b) Densitometric measurement of CKIε immunoreactivity from western blotting. Data represent percentages of the overall maximal value. Each value is the mean ± SEM (n = 3). (c) Double-labeling of vertical sections of rat retinae by confocal immunofluorescence for CKIε protein (sections i, iv, vii) and RIBEYE/CtBP2 (sections ii, v, viii). Anti-RIBEYE/CtBP2 stains synaptic ribbons (anti-B domain of RIBEYE) indicating the region of the photoreceptor ribbon synapses in the outer plexiform layer (OPL). In addition, it immunostains the envelope of cell nuclei of photoreceptors in the outer nuclear layer (ONL) and the entire cell nuclei of neurons in the inner nuclear layer (INL) (anti-CtBP2). In (iii), (vi) and (ix), images are merged to demonstrate that in cell somata immunoreactivity for CKIε is localized in the cytosol. Note that, in photoreceptors and neurons of the inner nuclear layer (INL), the abundance of CKIε is higher at ZT18 (iv, vi) and ZT0 (vii, ix) than at ZT6 (i, iii).

Discussion

The significance of the clock system in mammalian retina for retinal physiology and pathophysiology is widely known, although the participation and the functioning of the individual clocks forming the clock system remain largely unknown. This is attributable to the difficulties inherent in the monitoring of the clockwork mechanism in specific types of neurons under in vivo conditions. In the present study, by combining laser capture micro-dissection with quantitative PCR, a protocol has been developed that is applicable to the ex vivo isolation of the transcriptomes of photoreceptor cells and allows the investigation of the endogenous photoreceptor clockwork and its target gene Aa-nat.

In accordance with previous studies (Niki et al. 1998; Liu et al. 2004), we have observed that the amount of Aa-nat mRNA displays a circadian rhythm with elevated values in the photoreceptors at night. Because the Aa-nat gene represents a target gene of the circadian clock of the photoreceptor (for reviews, see Tosini and Fukuhara 2003; Iuvone et al. 2005), the presently observed cyclicity in the amount of Aa-nat transcript reflects the functioning of the photoreceptor clock in our experimental system. The rhythm of the amount of Aa-nat mRNA in photoreceptors is phase-delayed under DD. As Aa-nat expression is phase-advanced after removal of the inner retina (Sakamoto et al. 2006) the drop-out of a non-circadian clock in the inner retina under DD may account for this effect.

In rat photoreceptors, the clock genes Clock, Bmal1, Per1, Per3, Cry2, and CkIε display 24-h changes in the amount of transcript, changes that are retained for Clock and Per3 after not less than three cycles of DD. On the basis of these findings, mammalian photoreceptors are concluded to contain a circadian clock that functions not only in vitro in disconnected layers (Tosini et al. 2007a), but also in vivo in the intact retina (as demonstrated here). At first glance, the functioning of the photoreceptor clock in the intact retina contradicts previous studies reporting that bioluminescence fails to be rhythmic in cultured whole retinas of Period-luciferase rats (Tosini et al. 2007a) and mice (Ruan et al. 2006). The discrepancy can be explained by the fact that oscillation of Period transcripts appears to occur in more than one retinal layer (Ruan et al. 2006; Tosini et al. 2007a) with probably different phases (Tosini et al. 2007a). It is therefore likely that the bioluminescence measured in the whole retina represents an overall level of bioluminescence produced from different clocks that oscillate out of phase. Accordingly, the overall level of bioluminescence may be constant although the bioluminescence produced from the individual clocks is rhythmic.

Despite the different methods used in monitoring gene expression, the phasing of the cycles in Per1 expression is similar, regardless of whether photoreceptors are investigated in vitro in disconnected layers (Tosini et al. 2007a) or in vivo (as seen in this study). This result is in accordance with the concept of the photoreceptor clock being independent of the master clock in the SCN. Moreover, it suggests that the photoreceptor clock functions independently of the inner retina and is thus not set by dopamine or other neurotransmitters released from neurons in the inner retina. As basic photoreceptor functions that are under circadian control are regulated by the inner retina via dopamine (for review, see Iuvone et al. 2005) this suggestion speaks against a role of the photoreceptor clock in mediating the influence of dopamine on photoreceptor function. Interestingly, the clock in non-mammalian photoreceptors (Besharse et al. 2004) depends on dopamine, suggesting a differential influence of dopamine on the photoreceptor clock in mammals and non-mammals.

The clock of photoreceptors resembles the central clock in the SCN in being autonomous, capable of storing photoperiodic information (Rohleder et al. 2006), and insensitive to feeding patterns (Oishi et al. 2003). However, it comprises oscillation of Clock (as shown in this study) and, in this regard, conforms to peripheral clocks (for review, see DeBruyne 2008) and differs from the central clock in the SCN (Maywood et al. 2003), in which Clock neither shows rhythmicity nor is essential for the functioning of the clock (DeBruyne et al. 2006). Hence, the photoreceptor clock exemplifies the idea that the cycling of Clock is not restricted to slave clocks but may also occur in autonomous clocks. In photoreceptors, the cyclicity of Clock may be important not only for maintaining the oscillation of the clock machinery (DeBruyne et al. 2007), but also for enabling the clock to control the daily regulation of gene transcription. This is evident from the finding that, in mammalian photoreceptors, Clock is important for regulating the E-box-dependent transcription of the genes Aa-nat (Chen and Baler 2000) and nocturnin (Li et al. 2008).

As shown in the present study, in photoreceptors, the cycling of Clock and Bmal1 occurs in-phase with Per1 and Per3. As, in the clocks investigated so far, the expression of Clock and Bmal1 oscillates in anti-phase to Period (for review, see Ko and Takahashi 2006), it can be concluded that the machinery of the photoreceptor clock is unique with regard to the intrinsic rhythmic properties of its molecular components. Synchronous oscillation of Clock/Bmal1 and Period is of general interest, because it exemplifies the proposal that the functioning of a clock does not necessarily require the sequential transcription of the genes constituting the positive and negative feedback loops of the molecular clockwork mechanism. This also appears to be valid for clocks in other tissues and species (for review, see Hastings et al. 2008b). Thus, when the transcription of Clock is temporally misaligned in transgenic Drosophila by driving it with the Per promoter, little effect is observed on the cycles of CLOCK protein expression, and circadian behavioral cycles are maintained (Kim et al. 2002). Sequential transcriptional activity of CLOCK/BMAL1 and PERIOD is considered to be a prerequisite of clock function. This raises the question of how the photoreceptor clock can function when oscillation of Clock/Bmal1 transcript and Period transcript is in phase. There is increasing evidence that oscillatory post-transcriptional events integrate with transcriptional feedback loops to sustain circadian pacemaking (for review, see Hastings et al. 2008b). If this is also valid for photoreceptors, a time lag between CLOCK/BMAL1 and PERIOD may be introduced by differential post-transcriptional events such as differential post-translational modification or differential intracellular trafficking.

The clockwork mechanism in photoreceptors shows marked oscillation of CkIε at the transcript and protein levels. Many studies in various organisms and tissues have shown that the CkIε gene is constitutively expressed under LD and DD conditions (Takano et al. 2000; Ishida et al. 2001; Lincoln et al. 2002; for review, see Mehra et al. 2009). Therefore, the clockwork within photoreceptors is unique in possessing circadian changes in CkIε expression. Whereas the expression of CkIε is generally constitutive in SCN, the activity of CKIε has been reported to display circadian changes in SCN of hamster (Agostino et al. 2008). It is therefore tempting to speculate that changes in CKIεactivity may occur in both photoreceptors and SCN, with the difference that these changes are produced at the transcriptional level in photoreceptors (as seen in this study) and at the post-transcriptional level in SCN (for review, see Mehra et al. 2009). CKIε mediates the phosphorylation of clock gene proteins and thus influences their nuclear translocation. Therefore, the observed daily rhythm of CkIε in photoreceptors might ensure that the phosphorylation and nuclear translocation of clock components occur in a phase-specific manner.

In order to compare the clock of photoreceptors with other clocks of the retinal clock system, the expression of clock genes has also been recorded in preparations of whole retina under the same experimental conditions. In accordance with previous studies, the clock genes Clock (Rohleder et al. 2006; Tosini et al. 2007b), Bmal1 (Kamphuis et al. 2005), Per1 (Rohleder et al. 2006; Tosini et al. 2007b), Per2 (Kamphuis et al. 2005; Rohleder et al. 2006; Tosini et al. 2007b), and Per3 (Kamphuis et al. 2005; Rohleder et al. 2006) display daily changes in the amount of transcript in preparations of whole rat retina. This suggests that Per2 oscillation is more pronounced in preparations of whole retina (p = 0.003) than in photoreceptors (p = 0.22), whereas for Clock, Bmal1, Per1, Per3, Cry2, and CkIε, the situation is reversed. On the basis of these findings, we conclude that the clocks in photoreceptors and neurons of the inner retina differ in the intrinsic rhythmicity of their molecular components. The present observation that exclusively the clocks of the inner retina feature the cycling of Per2 is consistent with the restriction of rhythmic Per2::Luciferase (LUC) expression to the inner nuclear layer in cultured retinas of mouse (Ruan et al. 2008) and with the persistence of Per2 rhythmicity in photoreceptor-degenerate retinas of the mouse (Ruan et al. 2006) and rat (Tosini et al. 2007b). The differential intrinsic rhythmicity of the clocks in photoreceptors and neurons of the inner retina may reflect their different functional significances. Whereas the clock of photoreceptors appears to drive melatonin formation, that of dopaminergic amacrine cells in the inner retina is suggested to control dopamine release (for reviews, see Tosini and Fukuhara 2003; Iuvone et al. 2005).

In conclusion, the present study provides insight into the clockwork of mammalian photoreceptors. It reveals that the photoreceptor clock is interesting not only with respect to its functional role in retinal physiology and pathology, but also in its capacity as a model system for studying the machinery of circadian clocks. More generally, the present study suggests that, across a range of tissues, circadian clocks disagree with respect not only to Clock, but also in terms of the role of CkIε and the interaction of the positive and negative feedback loops of the molecular oscillator.

Acknowledgements

We thank Ms. Kristina Schäfer, Ms. Ursula Göringer-Struwe, Ms. Ilse von Graevenitz, and Annette de Cuvry for their excellent technical assistance. We also thank Ms. Bettina Wiechers-Schmied and Ms. Susanne Rometsch for secretarial help and Mr. Nils Rohleder for graphical work. The data contained in this study are included in the theses of Ms. Katja Schneider and Ms. Susanne Tippmann as a partial fulfillment of their medical doctorate degree at the Johannes Gutenberg University, Mainz.

Ancillary