Regulation of photoreceptor Per1 and Per2 by ... - Wiley Online Library

ª Federation of European Neuroscience Societies

European Journal of Neuroscience, Vol. 20, pp. 167–174, 2004

Regulation of photoreceptor Per1 and Per2 by light, dopamine and a circadian clock Joseph C. Besharse, Minhong Zhuang, Katie Freeman and Joseph Fogerty Department of Cell Biology, Neurobiology, and Anatomy, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226, USA Keywords: cAMP, ganglion cell, inner nuclear layer, quinpirole, retina, Xenopus laevis

Abstract In the Xenopus laevis retina, a principal model for retinal circadian organization, photoreceptors have all the properties of circadian oscillators. However, rhythmic oscillations of Per1 gene expression in the inner retina (but not photoreceptors) have been reported in mice with the suggestion that mice and frogs have a different retinal circadian organization. Although it is known that two period genes (xPer1 and xPer2) exhibit different temporal patterns of expression in the Xenopus retina, and that one (xPer2) is directly responsive to light and dopamine, it is not known whether this reflects the properties of period genes within photoreceptor oscillators or among distinct retinal cell populations. We addressed this by determining the cellular site of light and dopamine regulated xPer2 expression, and the diurnal expression of both xPer1 and xPer2 using in situ hybridization. Our data show that both xPer1 and xPer2 are expressed in most cell types in the retina, including inner nuclear neurons and ganglion cells. However, light and quinpirole, a dopamine agonist, increase xPer2 levels specifically in photoreceptors, and the effect of quinpirole, but not light, is blocked by pCPTcAMP. Furthermore, antiphasic diurnal expression of xPer1 and xPer2 also occurs in photoreceptors. Our analysis does not provide insight into the near constitutive expression of period genes in the inner retina, but supports a model in which light- and dopamine regulated-xPer2 and rhythmic xPer1 play critical roles in entrainment and circadian oscillations within photoreceptors.

Introduction Cellular clocks that produce endogenous timing cues through coupled negative and positive feedback loops of gene expression control circadian rhythms. Among vertebrates these regulatory loops are currently best understood for the suprachiasmatic nucleus (SCN) of mice (reviewed in Reppert & Weaver, 2001, 2002; Maywood et al., 2003). There the basic helix–loop–helix, PAS domain transcription factors, CLOCK and BMAL1, drive transcription of period (Per1, Per2, and Per3), cryptochrome (Cry1 and Cry2) and Rev-erba genes. In turn, expression of CRY protein, and to a lesser extent PER protein, regulates CLOCK ⁄ BMAL1 driven transcription negatively whereas REV-ERBa regulates Bmal1 transcription (Preitner et al., 2002). These regulatory loops drive endogenous circadian oscillations that result in rhythmic physiological outputs such as the behavioural activity rhythm in mice. Although period genes are critical elements of the clockwork, their specific function is not well defined. This is emphasized by the fact that a single period gene in Drosophila suffices in the negative limb of clockwork (Panda et al., 2002) whereas mammalian systems have three period homologues (Per1, Per2, and Per3), two of which (Per1, Per2) are essential for the SCN clockwork (Zheng et al., 1999, 2001; Shearman et al., 2000a; Bae et al., 2001). It seems likely that the three period homologues have distinct functions. Although three period homologues are generally modelled as components of a single cellular

Correspondence: Dr Joseph C. Besharse, as above. E-mail: [email protected] Received 24 March 2004, accepted 26 April 2004

doi:10.1111/j.1460-9568.2004.03479.x

oscillator, the extent to which they are coexpressed in the same clock cells has not been evaluated thoroughly. Given the widespread expression of circadian clock genes in general (Zylka et al., 1998; Hogenesch et al., 2003; Schibler et al., 2003), and the recent finding that rhythmic and acutely regulated Per gene expression can occur in different cell populations of the SCN (Hamada et al., 2001; Karatsoreos et al., 2004; Yan & Silver, 2004), it is important to determine the extent to which the acute regulation and circadian profiles of different period genes reflect their behaviour within the same or different cell populations. Within the retina, this takes on particular importance with the recent report that mPer1 exhibits a circadian profile in retinal amacrine cells but not in photoreceptors of mice (Witkovsky et al., 2003). Although the cellular organization of clock gene expression has not been extensively studied in any retinal system, these findings have led to the suggestion that retinas of mice and frogs differ in their circadian organization. In this study we have used retinas from the African clawed frog, Xenopus laevis, to study differential regulation of two period genes, xPer1 and xPer2 using in situ hybridization. The Xenopus system is ideal for analysis because robust circadian rhythms persist in the retina (Besharse & Iuvone, 1983) and in photoreceptor layers (Cahill & Besharse, 1993) in culture, and the layered organization of the many retinal cell types is well defined. Previous work has documented that photoreceptors are circadian clocks, the phase of which can be shifted by light and dopamine (Cahill & Besharse, 1991, 1993). The fact that light and dopamine also increase xPer2 mRNA led us to suggest that xPer2 plays a role in phase resetting (Steenhard & Besharse, 2000; Zhuang et al., 2000). However, this model was based entirely on Northern analysis of whole retinal RNA. The retinal cell types

168 J. C. Besharse et al. exhibiting either acute or diurnal changes in Per expression were not identified directly.

Materials and methods Animals Adult male Xenopus laevis were obtained from Nasco (Ft. Atkinson, WI, USA). The frogs were entrained at 21 C on a cyclic lighting schedule (LD) with lights on at Zeitgeber time O (ZT0) and off at ZT12 for at least 14 days before all experiments. In some experiments eyecups were prepared for immediate fixation in 4% paraformaldehyde in culture medium, or for RNA extraction as described previously by Green et al. (1995). Animals were killed by decapitation. The brain and spinal cord were then pithed prior to eye removal. Animal care and experimental protocols were performed in accordance with institutional and federal guidelines.

Eyecup culture Eyecups were prepared in room light during the 2-h period between ZT10 and ZT12, placed in a defined medium of balanced salts and amino acids as described previously (Cahill & Besharse, 1991; Steenhard & Besharse, 2000), and cultured overnight for in vitro experiments that were initiated during the period of ZT22 to ZT2.5 the following morning. Eyecups prepared in this way retain circadian clock properties with a circadian phase corresponding to intact animals on the same LD cycle (Besharse & Iuvone, 1983; Cahill & Besharse, 1991). Groups of five eyecups from separate animals were cultured in darkness overnight at room temperature (21 ± 1 C) in 10 mL medium in 60 mm culture dishes in a rotating incubator (60 r.p.m.) with constant gassing (95% O2 ⁄ 5% CO2). The following day cultures were exposed to light, darkness or drugs.

Light, pCPT-cAMP and quinpirole treatment The in vitro experiments began in darkness with a 30 min pretreatment with the membrane permeable cAMP analogue, 8-(4-chlorophenylthio)adenosine 3¢,5¢-cyclic monophosphate (pCPT-cAMP; SigmaRBI, St. Louis, MO, USA), at a final concentration of 1 mm. A total of 200 lL of a 50 mM solution was added to each dish. After 30 min, light or quinpirole (Sigma-RBI) treatment was initiated; 10 lL of a 100 lM solution of quinpirole was added per dish for a final concentration of 100 nm. Nuclease-free water was added to control dishes at the same time as drug additions. Light (3–5 · 10)4 W ⁄ cm2) was delivered to eyecups from a tungsten-halogen lamp (Oriel Corporation, Stamford, CT, USA) at constant voltage (12 V). Light intensity was measured with a calibrated IL 700 radiometer (International Light, Newburyport, MA, USA) at the plane of culture dishes. Dark control eyecups were maintained adjacent to light treated groups in gas-permeable film canisters. After 3 h, the retinas were collected and frozen on dry ice for Northern blot analysis or fixed as eyecups in 4% paraformaldehyde in culture medium for in situ hybridization. Dark-treated retinas for Northern analysis were collected quickly (within 3 min) in normal room light; dark-treated eyecups for in situ analysis were fixed under infrared illumination.

Northern blot analysis (xPer2, xActin) Total RNA from individual retinas was extracted using TRIZOL reagent (Invitrogen, San Diego, CA, USA) as described previously

(Zhuang et al., 2000). RNA was denatured for 5 min at 65 C, separated by electrophoresis in a 1% agarose formaldehyde gel, transferred to a Bright Star positively charged nylon membrane (Ambion, Austin, TX, USA) and cross-linked by UV light. Hybridization was carried out using Ultrahyb solution (Ambion) according to manufacturerı´s instruction. Blots were hybridized with antisense [32P]dCTP-labelled single-stranded DNA with an activity of 1 · 106 c.p.m. ⁄ mL at 42 C overnight. Probes were generated by asymmetric polymerase chain reaction (PCR; Zhuang et al., 2000) from nucleotides 575–1066 of the xPer2 sequence (GenBank AF199499) or nucleotides 171–411 of Xenopus beta-actin sequence (GenBank AF0791610). Phosphorimaging as described (Zhuang et al., 2000) was used for quantitative analysis.

3¢ RACE of xPer1 Our prior Northern analysis of xPer1 (Zhuang et al., 2000) utilized a probe generated from a 186 bp PCR product of the putative xPer1 sequence. To obtain a more complete xPer1 sequence suitable for generation of in situ hybridization probes, we conducted 3¢ RACE on total RNA extracted with Trizol reagent (Life Technologies, San Diego, CA, USA) from retina obtained at ZT1.5 when xPer1 expression was near its peak level (Zhuang et al., 2000). Total RNA was treated with DNase I, and the SuperScript II RNase H-Reverse Transcriptase kit (Invitrogen) was used to perform the reverse transcription reaction with a 3¢ RACE adapter (Ambion; FirstChoice RLM-RACE kit; 5¢GCGAGCACAGAATTAATACGACTCACTATAGGT12VN-3¢) as a primer. The Expand Long Template PCR system (Roche) was used to perform PCR with a gene-specific forward primer (bp 17–38 of GenBank, Acc# AF250547) and a 3¢ RACE outer primer (from Ambion’s FirstChoice RLM-RACE kit; 5¢- GCGAGCACAGAATTAATACGACT-3¢). The reaction was run on a Perkin Elmer 9700 PCR machine with the following profile: 93 C for 2 min followed by 10 cycles of (93 C, 10 s; 58 C, 30 s; 68 C, 6 min); 22 cycles of (93 C, 10 s; 58 C, 30 s; 68 C, 6 min + 20 s ⁄ cycle), and finished with 68 C for 7 min. Two 3’ RACE products of 3.8 kb and 4.8 kb were cloned into a pCR-XL TOPO vector (Invitrogen) and the 3.8 kb product was sequenced completely. This sequence (GenBank, AF250547) was found to contain an open reading frame 51% identical (67% similar) to human Per1 over its entire length beginning at amino acid 128, a 3¢-untranslated region and polyadenylation site. Another Xenopus cds in GenBank (AY340103), differing from our 3¢ RACE product only in its 3¢ sequence, encodes a shorter open reading frame lacking a C-terminal sequence, but does contain an additional 112-amino acid N-terminal sequence corresponding to human PER1 beginning at its start codon.

In situ hybridization For the quantitative analysis, 33P-labelled antisense and sense riboprobes were prepared by in vitro transcription as described previously (Green et al., 1995). For xPer1 a 1016 bp fragment (nt 1006–2022 of GenBank AF250547) of the xPer1 open reading frame was subcloned into pCRII-TOPO vector (Invitrogen) using nt 1006– 1025 as a forward primer and the complement of nt 2004–22 as a reverse primer. A 996-bp fragment of xPer2 (nt 1039–2034, AF199499) was prepared in a similar manner. For xPer2 we also used nonradioactive, digoxigenin-labelled riboprobes detected with anti-digoxigenin antibodies as described (Schaeren-Wiemers & Gerfin-Moser, 1993; Zhu et al., 2000).

ª 2004 Federation of European Neuroscience Societies, European Journal of Neuroscience, 20, 167–174

Regulation of Per1 and Per2 in photoreceptors 169 Thick paraffin sections (10 lm) were placed on glass slides. Multiple sections from eyecups from each treatment or experimental group were placed adjacent to each other on the same slide to ensure identical treatment in subsequent steps. Slides were de-paraffinized and hydrated (Green et al., 1995) with procedural improvements described in detail by Rausa et al. (1998). Briefly, this involved treatment with 4% paraformaldehyde in phosphate buffered saline (PBS) for 20 min, a wash in PBS, treatment with 20 lg ⁄ mL proteinase K for 7.5 min at room temperature, a wash in PBS, and another treatment with 4% paraformaldehyde followed by PBS (5 min each). Sections were then treated with 0.1 m triethanolamine (TEA) pH 8.0 for 5 min, and 0.25% acetic anhydride ⁄ 0.1 m TEA for 10 min, and washed in 2 · standard saline citrate (SSC) twice for 5 min each. The 33P-labelled probe was added at a final activity of 5 10 c.p.m. ⁄ lL to a hybridization solution containing 40% formamide, 0.48 m NaCl, 8 mm Tris pH 7.5, 1.6 mm EDTA, 0.8 · Denhardt’s solution, 0.4 mg ⁄ mL yeast tRNA, 80 lg ⁄ mL poly(A) and 10% dextran sulphate. The mixture was heated to 80 C for 5 min, and dithiothreitol added to a concentration of 50 mm from a 3 m stock. Approximately 50 lL of the probe solution was added to each slide and covered with a glass cover slip. Hybridizations were carried out overnight in a moist chamber equilibrated to 65 C containing 4 · SSC ⁄ 50% formamide. The next day, slides were washed in 5 · SSC for 10 min at 65 C, the coverslips were removed and the slides were washed with 50% formamide ⁄ 2 · SSC for 20 min at 65 C. At 37 C the slides were washed twice for 30 min in RNase buffer (0.3 m NaCl ⁄ 10 mm Tris pH 8 ⁄ 5 mm EDTA), treated with 50 lg ⁄ mL RNase in the same buffer for 30 min, and washed in the same buffer for 15 min at 37 C. Slides were then washed in 50% formamide ⁄ 2 · SSC for 20 min at 65 C, in 2 · SSC for 15 min and then in 0.1 · SSC for 15 min at room temperature. Finally, the slides were passed sequentially through 70% (twice), 95% and 100% ethanol for 10 min each and air-dried. The dried slides were coated with a thin layer of NTB-2 autoradiographic emulsion (Eastman-Kodak) diluted (1 : 1) with water (Green et al., 1995), and transferred to a light-tight slide box. Slides were exposed at 4 C for 5–28 days before developing with D19 developer (Eastman-Kodak). Slides were observed using phase contrast or Hoffman contrast microscopy without staining, or in some cases after staining with haematoxylin and eosin.

made from multiple sections of the same eyecup. With the aid of NIH Image software (Version 1.6.3), a rectangular field of known area was moved to multiple sites on each image to determine the included number of silver grains. Grains were counted over the inner segments of photoreceptors (IS), the inner nuclear layer (INL) at its junction with the inner plexiform layer, and the ganglion cell layer (GC). We also estimated background grain density over the vitreous adjacent to the retina. Background grain density, which was similar for slides hybridized with both sense and antisense probes, was subtracted from the grain density of each layer on an image-by-image basis. We used an equal number of images from two or three eyecups for all experimental groups. Data are presented as the mean number of silver grains per lm2 ± SD. Slides hybridized with sense probes did not have silver grains above background levels over the retina. However, we found that rod outer segments, which lack RNA, generally had silver grain densities exceeding background. This nonspecific labelling of outer segments occurred equally with both sense and antisense probes. Grain density data and quantitative Northern blot data were analysed using one-way analysis of variance with alpha at P ¼ 0.05 for multiple comparisons or with Student’s t-test for paired comparisons.

Results xPer2 mRNA is expressed in multiple cell types In order to determine which retinal cell types express xPer2 mRNA, we conducted in situ hybridization experiments using a digoxigenin detection procedure on samples prepared at the time of the diurnal peak (mid-morning) of xPer2 mRNA (Zhuang et al., 2000). xPer2 mRNA was abundant in both rod and cone photoreceptors (Fig. 1A). However, a hybridization signal at levels well above that seen with the sense control (Fig. 1B) was detected in the INL and GC layers. In particular, significant expression of xPer2 was detected in cells with an amacrine-like morphology at the margin of the INL and inner plexiform layer (IPL) and in all

Quantitative analysis 33

P-labelled riboprobes permitted both cellular localization with good resolution and quantitative analysis of xPer1 and xPer2 mRNA in retinal layers (Green et al., 1995; Rausa et al., 1998). Deterioration of resolution in tissue autoradiograms from scatter of b-electrons increases with increasing energy level (Salpeter et al., 1974); the average energy level of b-electrons from 33P is much lower than for 32 P and provides sufficient resolution to distinguish between adjacent rods and cones in the Xenopus retina (Green et al., 1995). Quantitative analysis involved direct counting of individual silver grains, which is directly related to the number of decay events so long as the emulsion is not saturated. We prepared 3–5 sets of slides for each experiment and developed them at periods ranging from 5 to 28 days. Only slide sets in which grain density on the most highly exposed cells was well below the saturation level of the emulsion were used for density analysis. Five full-thickness digital images of the central retinal field were made for each eyecup with Hoffman contrast optics; images were

Fig. 1. xPer2 in situ hybridization of the Xenopus laevis retina with the digoxigenin technique at ZT2.5 shows labelling over multiple cell types with the antisense (A) but not with the sense (B) probes. Photoreceptors (both rod and cone) and cells at the margin of the inner nuclear and inner plexiform layers (black arrow) and ganglion cell layers (white arrow) were prominently labelled. However, virtually all cells within the retina also exhibited significant labelling above background. RPE, retinal pigment epithelium; OS, outer segments; ON, outer nuclear layer; IN, inner nuclear layer; IPL, inner plexiform layer; GC, ganglion cell layer.


170 J. C. Besharse et al. cell bodies in the ganglion cell layer. This, along with our previous data (Zhuang et al., 2000) showing rhythmic expression of xPer2 mRNA in the retinal pigment epithelium-choroid, demonstrates that xPer2 is expressed in many cell types in the eye.

Light and quinpirole induction of xPer2 mRNA in photoreceptors Light and dopamine (or its analogue, quinpirole) acutely increase retinal xPer2, but not xPer1 mRNA in Xenopus laevis eyecups in vitro (Steenhard & Besharse, 2000). Here we extend this finding with data showing that the effect of quinpirole, but not light, is attenuated by inclusion of the cAMP analogue pCPT-cAMP (Fig. 2). Figure 2 illustrates two separate experiments with representative Northern blots (Fig. 2A and B) and quantitative analysis of xPer2 relative to xActin mRNA (Fig. 2C and D). Although the magnitude varies among experiments, light and qunipirole always increase xPer2 level (see also Steenhard & Besharse, 2000); the effects in the two experiments illustrated are highly significant (P ¼ 0.0001) statistically. Addition of pCPT-cAMP to the medium has no effect in light (Fig. 2A and C), but attenuates the effect of quinpirole in darkness (Fig. 2B and D). The blocking effect of pCPT-cAMP on quinpirole in darkness is statistically significant (P ¼ 0.005) and has been observed in three independent experiments (not shown). Nonetheless, the blocking effect of pCPT-cAMP may be confounded by the fact that pCPTcAMP also causes a modest but significant (P ¼ 0.005) increase in xPer2 mRNA when provided alone in darkness (see Fig. 2C). These differential effects of pCPT-cAMP could reflect the behaviour of different cell types within the retina (see below). To evaluate the effects of light, quinpirole and pCPT-cAMP in different cell populations, we conducted experiments similar to those shown in Fig. 2, but examined xPer2 levels by in situ hybridization (Fig. 3). xPer2 mRNA was present in many different cell types within the retina, but light (Fig. 3B) and quinpirole (Fig. 3E) increased xPer2 mRNA levels above dark levels (Fig. 3A) specifically in photoreceptors. Furthermore, the effect of quinpirole (Fig. 3F), but not light (Fig. 3C), in photoreceptors was blocked by inclusion of pCPT-cAMP. Quantitative analysis of silver grains (grains ⁄ lm2) over photoreceptor IS, over the inner nuclear where it borders with the INL and over GCs revealed (Fig. 4A) that both light and quinpirole significantly (P < 0.001) increased grain density (by as much as sixfold compared with the dark control) over photoreceptor IS. Furthermore, the effect of quinpirole (but not light) was greatly attenuated by addition of pCPT-cAMP. In the darkness, light, or darkness plus quinpirole groups, grain density over the INL or GC layers was low and significant differences among these groups were not detected. Compared with similar Northern blot data (Fig. 3C and D), the magnitude of the induction of xPer2 is 2–3-times higher and the apparent blocking effect of pCPT-cAMP is more substantial when the signal from photoreceptors is considered alone. Other than the effects of light, quinpirole and pCPT-cAMP specifically in photoreceptor IS, only one other significant effect was noted in these comparisons. Grain density increased about twofold specifically in the GC layer in all groups treated with pCPT-cAMP (Fig. 4A). This upregulation of xPer2 mRNA in the GC layer was statistically significant (P ¼ 0.05) for the comparison of darkness vs. darkness plus cAMP and quinpirole plus pCPTcAMP. This effect in ganglion cells in a direction opposite to that in photoreceptors could account for the small upregulation in the presence of pCPT-cAMP often seen (see Fig. 3C) in whole retinal extracts by Northern analysis.

Fig. 2. Northern analysis of the effects of light, quinpirole, and pCPT-cAMP on retinal xPer2 mRNA level. (A) xPer2 in darkness, light, light (L) + cAMP, and dark (D) + cAMP; each group includes three adjacent lanes representing three retinas from different animals all taken from the same X-ray film. B. xPer2 in darkness, light, quinpirole in darkness and quinpirole (Q) + cAMP in darkness; each group includes three adjacent lanes representing three retinas from different animals all taken from the same X-ray film. (C) Quantitative analysis using phosphor-imaging of the experiment illustrated in A with five retinas per treatment group. Relative mRNA is the ratio of xPer2 mRNA to xActin mRNA on the same blot. Data are standardized so that the dark group equals 1. Light and L + cAMP groups are different from darkness (P ¼ 0.0001); the D + cAMP group also differs (P ¼ 0.005) from darkness. (D) Quantitative analysis (as in C) of the experiment illustrated in B with five retinas per treatment group. Light and quinpirole groups are different from darkness (P ¼ 0.0001); the Q + cAMP group is significantly different (P ¼ 0.005) from quinpirole alone.

xPer1 mRNA is not regulated by light or quinpirole Previous Northern analysis using a short (186 bp) PCR fragment of a putative xPer1 cDNA suggested that, unlike xPer2, xPer1 was not regulated by light or quinpirole in the retina (Steenhard & Besharse, 2000; Zhuang et al., 2000). To generate probes appropriate for in situ


Regulation of Per1 and Per2 in photoreceptors 171

Fig. 3. In situ hybridization with 33P-labled RNA probes showing that light and quinpirole induce xPer2 mRNA specifically in photoreceptors and the effect of quinpirole is blocked by pCPT-cAMP. Shown are phase contrast images of unstained retinal sections from eyecups treated 3 h in darkness (A), light (B), light plus pCPT-cAMP (C), darkness plus pCPT-cAMP (D), darkness plus quinpirole (E), or darkness plus quinpirole and pCPT-cAMP (F). At this magnification black silver grains are small and are most readily seen over photoreceptor inner segments in B, C and E. OS, outer segments; IS, photoreceptor inner segments; IN, inner nuclear layer; GC, ganglion cell layer.

analysis, we used primers based on the 186 bp fragment to obtain and sequence a 3’ RACE product of xPer1 that included most of the open reading frame, 3’-untranslated region and polyadenylation site. This sequence (GenBank, AF250547) was found to encode a PER protein with high sequence similarity throughout its length to human PER1. Using 33P-labelled sense and antisense probes based on a 1016-bp segment from this sequence, we extended the quantitative in situ hybridization analysis to xPer1. xPer1 mRNA was found in photoreceptors as well as within cells of the inner nuclear and ganglion cell layers (Fig. 5). Hybridization occurred predominantly in association with cell bodies and was much less prominent over plexiform layers and the optic nerve (Fig. 5A). Major differences in the level of hybridization among different retinal layers was not readily detected; cells at the margin of the inner nuclear and outer plexiform layer (IN) and GC layer were prominently labelled (Fig. 5B and C). Quantitative analysis of grain density using sections from the same eyecups as described for the analysis of xPer2 (Fig. 4B) showed that xPer1 to be present at roughly equal levels in the IS, INL and GC compartments. Furthermore, there was no statistically significant (P ¼ 0.05) effect of light, quinpirole, or pCPTcAMP on xPer1 mRNA level in any retinal compartment (Fig. 4B).

Antiphasic rhythms of xPer1 and xPer2 mRNA in photoreceptors xPer1 mRNA exhibits a robust diurnal rhythm in LD and a circadian rhythm in DD, peaking near the time of light onset (ZT0), while xPer2 is at its lowest level at the time of light onset and peaks during the daytime in response to light (Zhuang et al., 2000). In order to determine which cells exhibit diurnal oscillations of xPer1 and xPer2 mRNA, we conducted a quantitative in situ hybridization analysis of eyes from animals maintained in LD and prepared near the peak (ZT0) and nadir (ZT8) of the xPer1 rhythm. Again, xPer1 was expressed in multiple cell types as described in the in vitro experiment (Fig. 4), but a

Fig. 4. Quantitative analysis of in situ hybridization sections for xPer2 (A) and xPer1 (B) mRNA from retinas fixed as in Figure 3. The Y-axis depicts silver grain density (grains ⁄ lm2) over photoreceptor inner segments (IS), cells along the border of the inner nuclear and inner plexiform layers (INL), and cells in the ganglion cell layer (GC). The X-axis depicts categories identical to the treatments illustrated in Figure 3.


172 J. C. Besharse et al.

Fig. 5. In situ hybridization with 33P-labelled RNA probes showing that xPer1 mRNA is found in multiple cell types throughout the retina. (A and B) Unstained sections imaged with Hoffman differential contrast optics. (C) Brightfield image of a haematoxylin and eosin-stained section. (A)Section through the optic nerve (ON) showing a low level of hybridization in the optic nerve and plexiform layers and a high level of hybridization over inner segments, inner nuclear layer and ganglion cell layer. The arrow in A indicates a red blood cell. (B and C) Show patches of hybridization corresponding to cell bodies in the INL and GC layer. Although silver grains are found over nuclei, most are found immediately around the periphery of the nuclei. Scale bars, 50 lm (A–C).

statistically significant difference (P ¼ 0.001) was detected only for photoreceptors IS (Fig. 6A). xPer1 mRNA was approximately fourfold higher in this compartment at ZT0 than at ZT8, indicating that a diurnal oscillation of xPer1 mRNA occurs within photoreceptors. In contrast, xPer2 mRNA was at a low level in all retinal compartments at ZT0 and increased significantly (P ¼ 0.001) in the IS compartment at ZT8 (Fig. 6B). We also detected a small apparent increase in xPer2 mRNA in the INL and GC compartments at ZT8. However, this difference was not statistically significant (P ¼ 0.05).

Discussion Three general roles have been proposed for PER proteins in vertebrates. First, similar to the role of Drosophila PER in negative feedback on both Per and Tim (Timeless) transcription (Reppert & Weaver, 2000; Panda et al., 2002), mammalian Per homologues are

generally placed in the negative limb of the clockwork. Although mammalian Per genes are activated transcriptionally by CLOCK– BMAL1 (see Maywood et al., 2003), PER protein appears to be a relatively weak negative regulator compared with the CRY proteins, which also play this role in vertebrates (Kume et al., 1999). Second, PER2 has been proposed as a positive regulator of the Bmal1 mRNA rhythm (Shearman et al., 2000b), which is antiphasic to Per rhythms (Oishi et al., 1998). Third, Per1 and Per2 mRNA in both suprachiasmatic nuclei (SCN; Shearman et al., 1997; Okamura et al., 1999) and retina (Namihira et al., 2001) are induced by light, which has led to the idea that these genes are important for regulation of circadian phase. The precise mechanism of phase regulation remains unclear, however, and depends critically on the cellular organization of the tissue. For example, light induction of PER proteins could influence circadian phase through direct interaction with other components of the core clockwork such as other Pers or Bmal1 if light induction


Regulation of Per1 and Per2 in photoreceptors 173

Fig. 6. Quantitative analysis of in situ hybridization sections for xPer2 (A) and xPer1 (B) mRNA in retinas fixed at ZT0 and at ZT8 (X-axis). The Yaxis depicts silver grain density (grains ⁄ lm2) over photoreceptor inner segments (IS), cells along the border of the INL and IPL and cells in the GC layer.

occurs within the cells responsible for the core oscillation. However, an alternative idea that would require cell–cell interaction is implied by recent reports that the rhythmic (circadian) and light induced features of Per1 and Per2 occur in different cell regions of the SCN (Hamada et al., 2001; Karatsoreos et al., 2004; Yan & Silver, 2004). Expression of xPer1 and xPer2 in the Xenopus laevis retina is unusual compared to mammalian circadian systems (Zhuang et al., 2000). First, light and dopamine or its agonist quinpirole acutely increase xPer2 mRNA; xPer1 is not increased by either treatment. In mammalian systems both Per1 and Per2 mRNA are induced by light (Shearman et al., 1997; Okamura et al., 1999; Namihira et al., 2001). Second, xPer1 mRNA exhibits a circadian pattern of expression, rising during the latter part of the dark period and peaking near the time of light onset, while xPer2 exhibits a light-dependent, diurnal rhythm with a low point near the time of light onset and a peak during the daytime. Thus, in a light–dark cycle xPer1 and xPer2 have nearly antiphasic rhythms. In contrast, in mouse both mPer1 and mPer2 exhibit circadian expression patterns with similar circadian phases (Zylka et al., 1998). To date studies of Per gene expression in Xenopus has been limited to the mRNA level, and the lack of suitable antibodies led us in this study to focus on localization of mRNA. However, recent data (J.C.B., unpublished) with a new xPER2 antibody indicates that xPer2 mRNA levels accurately predict xPER2 protein levels.

Based on earlier work demonstrating phase resetting of the photoreceptor circadian clock by light or dopamine (Cahill & Besharse, 1993), we proposed that light and dopamine inducible xPer2 in photoreceptors was directly involved in resetting of the photoreceptor clock, an idea consistent with the fact that similar timing and doses of light or dopamine caused both circadian phase shifts (Cahill & Besharse, 1993) and increases in xPer2 mRNA (Steenhard & Besharse, 2000). In the current study, the relationship between phase shifting and xPer2 induction is further supported by the finding that pCPT-cAMP inhibits the effect of quinpirole (but not light) on xPer2 induction (Figs 2–4). This treatment was added to our experimental paradigm because of the recent demonstration that pCPT-cAMP inhibits the phase shifting effects of dopamine and quinpirole but not that of light (Hasegawa & Cahill, 1999). The xPer2 gene might be the point of convergence of the separate signalling pathways for light and dopamine on the photoreceptor clock proposed previously to account for the separate effects of light and dopamine on circadian phase (Cahill & Besharse, 1991). We also proposed that light or dopamine induced increases in xPER2 protein could influence circadian phase by altering expression of other clock components such as xPer1. For example, rising levels of xPer1 during the early evening could be delayed by increases in lightinduced xPER2 protein through negative feedback at the xPer1 promoter. This would lead to a delay phase-shift like that demonstrated to occur in the early evening (Cahill & Besharse, 1991). Late night induction of xPer2 could lead to advance phase shifts through similar mechanisms. The implicit assumption of this thinking was that xPer1 and xPer2 were coexpressed in photoreceptors. However, given the widespread expression of circadian clock genes (Zylka et al., 1998; Hogenesch et al., 2003; Schibler et al., 2003), and the recent report that mPer1 is expressed in a circadian pattern in amacrine cells in the mouse (Witkovsky et al., 2003), alternative models involving differential expression of xPer1 and xPer2 in different cell types are also possible. The new in situ hybridization data reported here show that whereas xPer1 and xPer2 mRNA are expressed in multiple cell types in the Xenopus laevis retina, both the light and quinpirole induction of xPer2 and the antiphasic diurnal oscillations of xPer1 and xPer2 reported previously (Steenhard & Besharse, 2000; Zhuang et al., 2000) reflect changes within photoreceptors. As the products of other central clock genes such as Clock (Zhu et al., 2000) and three Crys (Zhu & Green, 2001) have been localized to photoreceptors, and overexpression of a dominant negative CLOCK protein within photoreceptors ablates circadian output of melatonin (Hayasaka et al., 2002), it appears that the photoreceptor clock mechanism in Xenopus involves the same group of central clock genes that have been identified in mammalian systems. A principal difference, however, lies in the separate roles of xPer1 and xPer2 reflected in their antiphasic relationship. Further analysis of xPer1 and xPer2 at the molecular and cellular level in the Xenopus eye will likely provide insight into the mechanisms underlying differential expression of the two period genes and the mechanism by which acute changes in xPer2 regulate circadian phase. In addition to the well-documented photoreceptor clock mechanism, there is now substantial evidence supporting the concept of an inner retinal clock (reviewed in Green & Besharse, 2004). It seems likely that the relative role of inner retinal and photoreceptor based oscillators could vary among different species. For example, it has been suggested that the tightly coupled, antiphasic rhythms of melatonin from photoreceptors and dopamine from the inner retina of pigeons reflects the activity of independent oscillators (Adachi et al., 1999). Recent studies in rodents showing rhythmic expression of mPer1 in dopaminergic amacrine cells (Witkovsky et al., 2003),


174 J. C. Besharse et al. and continuation of a circadian rhythm in dopamine metabolism after degeneration of photoreceptors (Doyle et al., 2002), both suggest an inner retinal oscillator. Although we cannot rule out the existence of a low amplitude rhythm in the inner retina, out of phase with that in photoreceptors, the near constitutive behaviour of xPer mRNAs in the inner retina is similar to the recent finding of subpopulations of SCN neurons in which Per mRNA was not regulated rhythmically (Hamada et al., 2001), and might reflect a lesser role for inner retinal Per gene expression in the Xenopus retina, which is dominated by photoreceptor-based oscillators.

Acknowledgements The authors thank Phillip Wang for technical assistance and Carla Green and David Lourim for a critical reading of the manuscript. National Institutes of Health Research Grant EY02414 supported this work.

Abbreviations GC, ganglion cell; IL, inner layer; INL, inner nuclear layer; IPL, inner plexiform layer; PCR, polymerase chain reaction; SSC, standard saline citrate; ZT, Zeitgeber time.

References Adachi, A., Suzuki, Y., Nogi, T. & Ebihara, S. (1999) The relationship between ocular melatonin and dopamine rhythms in the pigeon: effects of melatonin inhibition on dopamine release. Brain Res., 815, 435–440. Bae, K., Jin, X., Maywood, E.S., Hastings, M.H., Reppert, S.M. & Weaver, D.R. (2001) Differential functions of mPer1, mPer2, and mPer3 in the SCN circadian clock. Neuron, 30, 525–536. Besharse, J.C. & Iuvone, P.M. (1983) Circadian clock in Xenopus eye controlling retinal serotonin N-acetyltransferase. Nature, 305, 133–135. Cahill, G.M. & Besharse, J.C. (1991) Resetting the circadian clock in cultured Xenopus eyecups: regulation of retinal melatonin rhythms by light and D2 dopamine receptors. J. Neurosci., 11, 2959–2971. Cahill, G.M. & Besharse, J.C. (1993) Circadian clock functions localized in Xenopus retinal photoreceptors. Neuron, 10, 573–577. Doyle, S.E., McIvor, W.E. & Menaker, M. (2002) Circadian rhythmicity in dopamine content of mammalian retina: role of the photoreceptors. J. Neurochem., 83, 211–219. Green, C.B. & Besharse, J.C. (2004) Retinal circadian clocks and control of retinal physiology. J. Biol. Rhythms, 19, 91–102. Green, C.B., Cahill, G.M. & Besharse, J.C. (1995) Tryptophan hydroxylase is expressed by photoreceptors in Xenopus laevis retina. Vis. Neurosci., 12, 663–670. Hamada, T., LeSauter, J., Venuti, J.M. & Silver, R. (2001) Expression of Period genes: rhythmic and nonrhythmic compartments of the suprachiasmatic nucleus pacemaker. J. Neurosci., 21, 7742–7750. Hasegawa, M. & Cahill, G.M. (1999) A role for cyclic AMP in entrainment of the circadian oscillator in Xenopus retinal photoreceptors by dopamine but not by light. J. Neurochem., 72, 1812–1820. Hayasaka, N., LaRue, S.I. & Green, C.B. (2002) In vivo disruption of Xenopus CLOCK in the retinal photoreceptor cells abolishes circadian melatonin rhythmicity without affecting its production levels. J. Neurosci., 22, 1600– 1607. Hogenesch, J.B., Panda, S., Kay, S. & Takahashi, J.S. (2003) Circadian transcriptional output in the SCN and liver of the mouse. Novartis Found. Symp., 253, 171–180. [Discussion 152–175, 102–179, 180–173 passim.] Karatsoreos, I.N., Yan, L., LeSauter, J. & Silver, R. (2004) Phenotype matters: identification of light-responsive cells in the mouse suprachiasmatic nucleus. J. Neurosci., 24, 68–75. Kume, K., Zylka, M.J., Sriram, S., Shearman, L.P., Weaver, D.R., Jin, X., Maywood, E.S., Hastings, M.S. & Reppert, S.M. (1999) MCRY1 and mCRY2 are essential components of the negative limb of the circadian clock feedback loop. Cell, 98, 193–205. Maywood, E.S., O’Brien, J.A. & Hastings, M.H. (2003) Expression of mCLOCK and other circadian clock-relevant proteins in the mouse suprachiasmatic nuclei. J. Neuroendocrinol., 15, 329–334.

Namihira, M., Honma, S., Abe, H., Masubuchi, S., Ikeda, M. & Honma, K. (2001) Circadian pattern, light responsiveness and localization of rPer1 and rPer2 gene expression in the rat retina. Neuroreport, 12, 471–475. Oishi, K., Sakamoto, K., Okada, T., Nagase, T. & Ishida, N. (1998) Antiphase circadian expression between BMAL1 and period homologue mRNA in the suprachiasmatic nucleus and peripheral tissues of rats. Biochem. Biophys. Res. Commun, 253, 199–203. Okamura, H., Miyake, S., Sumi, Y., Yamaguchi, S., Yasui, A., Muijtjens, M., Hoeijmakers, J.H.J. & van der Horst, J.H.J. (1999) Photic induction of mPer1 and mPer2 in Cry-deficient mice lacking a biological clock. Science, 286, 2531–2534. Panda, S., Hogenesch, J.B. & Kay, S.A. (2002) Circadian rhythms from flies to human. Nature, 417, 329–335. Preitner, N., Damiola, F., Lopez-Molina, L., Zakany, J., Duboule, D., Albrecht, U. & Schibler, U. (2002) The orphan nuclear receptor REV-ERBalpha controls circadian transcription within the positive limb of the mammalian circadian oscillator. Cell, 110, 251–260. Rausa, F.M.,Ye, H., Lim, L., Duncan, S.A. & Costa, R.H. (1998) In situ hybridization with 33P-labeled RNA probes for determination of cellular expression patterns of liver transcription factors in mouse embryos. Methods, 16, 29–41. Reppert, S.M. & Weaver, D.R. (2000) Comparing clockworks: mouse versus fly. J. Biol. Rhythms, 15, 357–364. Reppert, S.M. & Weaver, D.R. (2001) Molecular analysis of mammalian circadian rhythms. Annu. Rev. Physiol., 63, 647–676. Reppert, S.M. & Weaver, D.R. (2002) Coordination of circadian timing in mammals. Nature, 418, 935–941. Salpeter, M.M., Budd, G.C. & Mattimoe, S. (1974) Resolution in autoradiography using semithin sections. J. Histochem. Cytochem., 22, 217–222. Schaeren-Wiemers, N. & Gerfin-Moser, A. (1993) A single protocol to detect transcripts of various types and expression levels in neural tissue and cultured cells: in situ hybridization using digoxigenin-labelled cRNA probes. Histochemistry, 100, 431–440. Schibler, U., Ripperger, J. & Brown, S.A. (2003) Peripheral circadian oscillators in mammals: Time and food. J. Biol. Rhythms, 18, 250–260. Shearman, L.P., Jin, X., Lee, C., Reppert, S.M. & Weaver, D.R. (2000a) Targeted disruption of the mPer3 gene: subtle effects on circadian clock function. Mol. Cell Biol., 20, 6269–6275. Shearman, L.P., Sriram, S., Weaver, D.R., Maywood, E.S., Chaves, I., Zheng, B., Kume, K., Lee, C.C., van der Horst, G.T.J., Hastings, M.H. & Reppert, S.M. (2000b) Interacting molecular loops in the mammalian circadian clock. Science, 288, 1013–1018. Shearman, L.P., Zylka, M.J., Weaver, D.R., Kolakowski, L.F. Jr & Reppert, S.M. (1997) Two period homologs: Circadian expression and photic regulation in the suprachiasmatic nuclei. Neuron, 19, 1261–1269. Steenhard, B.M. & Besharse, J.C. (2000) Phase shifting the retinal circadian clock: xPer2 mRNA induction by light and dopamine. J. Neurosci., 20, 8572–8577. Witkovsky, P., Veisenberger, E., LeSauter, J., Yan, L., Johnson, M., Zhang, D.Q., McMahon, D. & Silver, R. (2003) Cellular location and circadian rhythm of expression of the biological clock gene Period 1 in the mouse retina. J. Neurosci., 23, 7670–7676. Yan, L. & Silver, R. (2004) Resetting the brain clock: time course and localization of mPER1 and mPER2 protein expression in suprachiasmatic nuclei during phase shifts. Eur. J. Neurosci., 19, 1105–1109. Zheng, B., Albrecht, U., Kaasik, K., Sage, M., Lu, W., Vaishnav, S., Li, Q., Sun, Z.S., Eichele, G., Bradley, A. & Lee, C.C. (2001) Nonredundant roles of the mPer1 and mPer2 genes in the mammalian circadian clock. Cell, 105, 683–694. Zheng, B., Larkin, D.W., Albrecht, U., Sun, Z.S., Sage, M., Eichele, G., Lee, C.C. & Bradley, A. (1999) The mPer2 gene encodes a functional component of the mammalian circadian clock. Nature, 400, 169–173. Zhu, H. & Green, C.B. (2001) Three cryptochromes are rhythmically expressed in Xenopus laevis retinal photoreceptors. Mol. Vis., 7, 210–215. Zhu, H., LaRue, S., Whitely, A., Steeves, T.D., Takahashi, J.S. & Green, C.B. (2000) The Xenopus clock gene is constitutively expressed in retinal photoreceptors. Mol. Brain Res., 75, 303–308. Zhuang, M., Wang, Y., Steenhard, B.M. & Besharse, J.C. (2000) Differential regulation of two period genes in the Xenopus eye. Brain Res. Mol. Brain Res., 82, 52–64. Zylka, M.J., Shearman, L.P., Weaver, D.R. & Reppert, S.M. (1998) Three period homologs in mammals: Differential light responses in the suprachiasmatic circadian clock and oscillating transcripts outside the brain. Neuron, 20, 1103–1110.
