In Situ Labeling of Transcription Sites in Marine Medaka

Abstract

Transcription factories have been characterized in cultured mammalian cells, but little is known about the regulation of these nuclear structures in different primary cell types. Using marine medaka, we observed transcription sites labeled by the metabolic incorporation of 5-fluorouridine (5-FU) into nascent RNA. Medaka was permeable to 5-FU in ambient water and became fully labeled within 4 hr of incubation. The incorporation of 5-FU was inhibited by the transcription inhibitor actinomycin D. The 5-FU incorporation sites were detected in the cell nucleus, and could be abolished by RNase digestion. The tissue distribution of 5-FU incorporation was visualized by immunocytochemistry on whole-mount specimens and histological sections. The 5-FU labeling appeared highly cell type specific, suggesting a regulation of the overall transcription activities at tissue level. Mapping of transcription factories by 5-FU incorporation in fish provides a useful and physiologically relevant model for studying the control of gene expression in the context of the functional organization of the cell nucleus. This manuscript contains online supplemental material at http://www.jhc.org. Please visit this article online to view these materials. (J Histochem Cytochem 58:173–181, 2010)

The intranuclear organization of transcription events has been a topic of intensive research in the past four decades. Transcription mediated by RNA polymerase I is localized in specific nucleolar sites (Lam et al. 2005). Likewise, increasing evidence indicates that transcription catalyzed by polymerases II and III is also structurally organized in thousands of nucleoplasmic foci known as transcription factories (reviewed by Carter et al. 2008). The control of the number and precise intranuclear locations of these factories, and the dynamic targeting of specific gene loci to these sites, are some of the most fundamental questions in cell biology.

Transcription factories have been mapped in mammalian cells by metabolic labeling of nascent RNA tethered on the sites of synthesis. Traditionally, transcription sites are labeled by in-situ nuclear run-on assays using halogenated nucleoside triphosphates, such as 5-bromouridine 5′-triphosphate (Br-UTP) (Iborra et al. 1996). The labeled RNA can be detected immunocytochemically by antibodies raised against BrdU that cross-react with the Br moiety of Br-UTP. Because Br-UTP is not cell permeable, this method can only be applied on mildly permeabilized, and therefore nonphysiological, cells. Alternatively, transcription can be pulse labeled by [³H]-uridine, and the localization of nascent RNA was revealed by autoradiography and electron microscopy (e.g., Wassermann et al. 1988). Uridine is not the immediate precursor for RNA synthesis; it is taken up by cells through specific transporters (Dahlig-Harley et al. 1981) and is phosphorylated intracellularly (Vidair and Rubin 2005) into UTP, which is then incorporated into RNA. More recently, a non-radioactive alternative to this method based on the use of epitope-tagged uridine, such as 5-fluorouridine (5-FU) or 5-bromouridine (5-BrU), has become popular in studying transcription factories (Boisvert et al. 2001; Trentani et al. 2003). Like [³H]-uridine, halogenated uridines are cell permeable, making it possible to incorporate the halogen tag in nascent RNA in living cells.

Despite the importance of transcription factories, virtually all information on these structures has been based on experiments using cultured mammalian cells. Little is known about the transcription factories in primary tissue cells. Casafont et al. (2006) reported the intranuclear distribution of nascent RNA in sensory ganglia neurons of rats injected intraperitonally with 5-FU. Similarly, Jao and Salic (2008) injected mice with 5-ethynyluridine and assayed transcription rates of various tissues in whole animals. These two studies demonstrate that it is feasible to metabolically label transcription factories in organisms. It is, however, still not clear how the number and morphology of transcription sites correlate with cell types, differentiation states, and functional activities of primary tissues.

In this study, we establish the medaka fish (Oryzias melastigma) as the model organism for in vivo transcription labeling. The medaka is generally permeable to exogenous molecules in ambient water (Hardman et al. 2008), making it a good candidate for metabolic labeling studies. The fish has been used extensively as the model for human genetic diseases (Dodd et al. 2000), and the nuclear architecture of fish is similar to that of mammals (e.g., Paris-Palacios and Biagianti-Risbourg 2006). By direct incubation of juvenile medaka with 5-FU in water, we observed extensive incorporation of 5-FU in all cell types in the fish, but the level of labeling varied according to cell type.

Materials and Methods

Fish Culture

Experiments were performed on 1-week-old medaka (Oryzias m.). The fish and embryos were collected from a laboratory colony and reared in 28‰ artificial sea water, which was prepared from deep sea synthetic sea salt (Lou Ton Aquarium; Hong Kong, China), at 27C with a 12:12-hr light/dark cycle.

Toxicity Analysis of 5-FU

Medaka were exposed to 5-FU (Sigma Aldrich; St. Louis, MO) at a concentration of 5 mM in 2 ml of sea water for 48 hr and observed for mortality every 6–12 hr. Medaka embryos were also exposed at the same concentration and duration, and observed for mortality and developmental defects every 24 hr. Each embryo was treated individually in a well of a 24-well plate. Images were captured with a digital color camera (Colorview 12, Olympus; Tokyo, Japan).

Histological and Immunohistochemical Analysis

Medaka were transferred to 2 ml of sea water containing 5 mM 5-FU and incubated for 4 hr before immunohistochemical analysis. All fish were anesthetized by transferring to aluminum foil and chilling on ice and were then snap-frozen in liquid nitrogen. The frozen fish were embedded with Tissue-Tek (Sakura; Torrance, CA) prior to cryosectioning.

Cryosections (10-μm) were fixed with 4% paraformaldehyde (PFA) for 5 min, rinsed three times with 1× phosphate-buffered saline (PBS), and permeabilized with 0.5% Triton X-100 in PBS for 5 min. After three washes with PBS, 5-FU incorporation was detected using monoclonal anti-BrdU (Sigma Aldrich) as the primary antibody (1:100) and goat anti-mouse IgG antibody (Sigma Aldrich) conjugated with tetramethyl rhodamine iso-thiocyanate (TRITC) as the secondary antibody (1:200). Hoechst 33342 (Sigma Aldrich) at 4 μM for 15 min was used as DNA counterstain. All slides were mounted with fluorescent mounting medium (DAKO; Carpinteria, CA) and sealed with nail polish.

To aid histological analysis of 5-FU incorporation, consecutive serial sections of medaka were stained with hematoxylin (EMS; Hatfield, PA) for 1 min. This was followed by a differentiation step, which was performed with 1% acid ethanol for 1–2 sec and bluing with 0.2% ammonia water for 1 min. Afterward, the section was counterstained with 0.2% eosin (EMS) in ethanol and serially dehydrated with ethanol. A clearing step with xylene was then performed to facilitate mounting the stained section with Permount mounting medium (Fisher; Pittsburgh, PA).

To investigate the FU staining pattern observed in medaka retina, an antibody marker for cone receptors was used. After 5-FU incubation, the medaka fish was frozen and serial-sectioned. One section was stained for 5-FU labeling as described above; the other was stained with zpr-1 monoclonal antibody (1:100, Oregon monoclonal bank).

In some experiments, whole-mount immunofluorescence was performed as modified from Chen et al. 2009. After 5-FU incubation, medaka was fixed with 4% PFA in PBS (4C, 2 hr) and was then serially dehydrated via methanol in PBS with 1% Tween 20 (USB; Cleveland, Ohio). The fish was then rehydrated by reversing the dehydration steps. Permeablization was conducted by incubating the fish in PBS with 1% Triton X-100, 1% DMSO, and 3% H₂O₂ for 2 hr. The specimen was blocked in 5% goat serum in PBS with 1% Tween 20. 5-FU incorporation was detected using monoclonal anti-BrdU as the primary antibody (1:100) and goat anti-mouse IgG antibody conjugated with TRITC as the secondary antibody (1:200). Hoechst 33342 at 4 μM for 1 hr was used as DNA counterstain. The medaka was then transferred to a glass coverslip and was covered by a drop of mounting medium. The coverslip was mounted on a microscope for imaging.

RNase A and Actinomycin D Treatment

To investigate whether 5-FU detection was RNA dependent, 5-FU–labeled medaka cryosections were fixed and permabilized, and then treated with 50 mg/ml RNase A from bovine pancreas (Sigma Aldrich) for 30 min prior to primary and secondary antibody incubation. To investigate whether 5-FU labels nascent RNA, medaka were treated with 10 mM actinomycin D (Sigma Aldrich) for 16 hr before 5-FU labeling as before.

Co-labeling 5-FU With 5-Ethyl-2′-deoxyuridine (E-dU)

For E-dU in vivo DNA labeling, 1 mM Ed-U (Invitrogen; Carlsbad, CA) was used to treat medaka for 12 hr prior 5-FU exposure. The medaka was then frozen, cryosectioned, and stained for 5-FU as described above, followed by Ed-U detection with the Click-It Ed-U kit (Invitrogen) according to the manufacturer's instructions.

Microscopy and Image Processing

Figures 1–5(except Figures 4A–4D) and Supplementary Figure 1 were collected using the Leica TCS SPE spectral confocal microscope (Leica Microsystems; Wetzlar, Germany) or the Leica TCS SP5 multiphoton confocal microscope (Leica Microsystems). Figures 4A–4D were imaged by a Deltavision Core widefield devolution microscope (Applied Precision; Issaquah, WA). Supplementary Figure 2 was collected by a Zeiss Axioplan microscope.

Figure 1.

5-Fluorouridine (5-FU) incorporation in medaka. (A) Hematoxylin and eosin staining of a sagittal section of medaka labeled with 5-FU. (B) 5-FU incorporated in medaka was detected by immunocytochemistry using anti-BrdU antibody (TRITC, red). (C) Same section as in B, counterstained by Hoechst 33342. (D–F) Distribution of 5-FU incorporation in the retina (D), brain (E), gut (F), and skin. (G) Organs were identified using the H and E–stained section (A) as a reference. (H–K) Corresponding Hoechst 33342 staining of D–F, respectively. (L–N) Subcellular locations of 5-FU incorporation in retinal cells. Transverse section of the medaka eye (L) was imaged at increasing magnifications (M,N), showing the punctate localization pattern of 5-FU. Bars: A–C = 200 μm; D–L = 50 μm; M,N = 5 μm.

Figure 2.

5-FU incorporation signals were RNA dependent. (A–C) 5-FU staining appeared as bright intranuclear foci (red) on a medaka retina section. (D–F) 5-FU–specific staining was absent when no 5-FU was used. (G–I) RNase A was used to treat sections from 5-FU–labeled medaka before immunohistochemical analysis. Bar = 10 μm.

Figure 3.

Simultaneous labeling of newly synthesized RNA and DNA in medaka. Sections of the retina (A–D), gut (E–H), and brain (I–L) of medaka colabeled with 5-FU and 5-ethyl-2′-deoxyuridine (E-dU). 5-FU (A,E,I; red) and E-dU (B,F,J; green) were incorporated into independent groups of cells. (M–P) Close-up of brain section showing the subcellular distribution patterns of 5-FU and E-dU incorporations. Bar = 50 μm.

Figure 4.

Whole-mount imaging of 5-FU incorporation in medaka. (A) 5-FU–treated medaka was labeled by whole-mount immunofluorescence. Red, 5-FU labeling; blue, Hoechst 33342 counterstaining. (B–D) Tail fin of the labeled fish was imaged at increasing magnifications. (E–H) Multiphoton confocal images of 5-FU incorporation collected at different depths into the medaka body. Bars: A = 200 μm; B = 20 μm; C,D = 2 μm; E–H = 50 μm.

Figure 5.

Cell type–specific transcription activities in medaka retina. (A–C) Transverse section of 5-FU–treated medaka labeled by anti-BrdU. (D–F) Serial section of the same retina labeled by anti-ZPR1. (G) A coronal section of medaka retina was labeled for 5-FU and imaged by multiphoton confocal sectioning from top (lens side) to bottom (pigmented epithelium side). The positions of 5-FU–positive cells were color coded (blue: top; red: bottom). (H–J) Maximum projection images of G. Bars: A–J = 50 μm; G = 10 × 8.3 μm.

All confocal images were digitally colored by the Leica LAS AF software (Leica Microsystems). Figures 4A–4D were processed from raw images using the deconvolution algorithm of the Deltavision Softworx software (Applied Precision). Figure 4A is a composite image montaged manually from 22 microscopy images made from the same fish. Edges of some of the images were removed by Photoshop (Adobe Systems; San Jose, CA). Color-coded three-dimensional reconstruction of confocal sections (Figure 5G) was done using Leica LAS AF.

Results

5-FU Treatment of Marine Medaka

Transcription sites in marine medaka could be labeled by a direct incubation of 1-week-old juvenile medaka with 5 mM 5-FU dissolved in 2 ml of sea water. The incorporated 5-FU was detected by immunocytochemistry (Boisvert et al. 2000) on cryosections of the fish after a mild fixation. In Figure 1, sagittal sections of the labeled medaka show that 5-FU staining was clearly visible in all nuclei (Figures 1B and 1C), suggesting that 5-FU could be effectively taken up and could permeate throughout the body of medaka. The labeling of 5-FU incorporation was evident after a 1-hr exposure and increased over time, reaching a plateau at 4 hr (data not shown). Although the 5-FU signal increased over time, the tissue distribution and intracellular pattern of labeling did not change, suggesting that the entry and biodistribution of 5-FU were not the limiting factors of incorporation. Although short pulses of 5-FU allow the efficient labeling of nascent RNA, 5-FU inhibits RNA modification (Samuelsson 1991) and splicing (Sierakowska et al. 1989), leading to cytotoxicity on prolonged exposure (Heimer and Sartorelli 1989). However, no mortality was observed in juvenile medaka exposed to 5 mM 5-FU for as long as 12 hr (data not shown). Fertilized medaka eggs, incubated with 5 mM 5-FU for 48 hr also appeared to develop normally (Supplementary Figure 1). The 5-FU toxicity was tested in time windows considerably longer than the duration of our metabolic labeling experiments. Taken together, the data show that 5-FU penetrated throughout the body of 1-week-old medaka and did not exhibit noticeable toxicity on prolonged exposure.

Tissue Distribution of 5-FU Incorporation in Medaka

An examination of 5-FU incorporation at high magnification revealed that 5-FU labeling was evident in all tissues and that the staining was restricted to the cell nucleus, consistent with the nascent RNA localization pattern. Significantly, the signal intensity was not the same for all cells. High-magnification views of the eye (Figure 1D), brain (Figure 1E), and liver (Figure 1F) show a mosaic staining pattern, suggesting that within the same organ, different cells may display different levels of 5-FU incorporation. Within individual cells, 5-FU labeled hundreds of discrete intranuclear foci (Figures 1L–1N), reminiscent of the transcription factories observed in mammalian cultured cells.

To confirm that 5-FU was incorporated into RNA, sections from 5-FU–treated medaka were digested with RNase A at a concentration commonly used to hydrolyze all cellular RNA (Krishan 1975). The immunostaining signal of the RNase-digested section was indistinguishable from that of the negative control (Figures 2D–2F), indicating that the nuclear 5-FU staining was totally abolished (Figures 2G–2I), and that the 5-FU staining was RNA dependent. Finally, prior treatment of medaka with actinomycin D, a toxin known to inhibit RNA synthesis (Sobell 1985), significantly reduced the signal of 5-FU incorporation (Supplementary Figure 2), consistent with the incorporation of 5-FU into nascent RNA.

Simultaneous Detection of Newly Synthesized RNA and DNA in Medaka

The cell-specific levels of 5-FU incorporation in medaka implied that transcription activities in different cells could be highly distinct. However, the tissue distribution of 5-FU incorporation might be complicated by other factors, such as the rates of cellular uptake and phosphorylation of 5-FU. To test whether the observed 5-FU incorporation pattern was a result of nucleoside biodistribution, we utilized another nucleoside analog, E-dU, that is converted by similar transport and processing pathways into a substrate for DNA replication. The incorporated E-dU can be detected by a fluorescently conjugated azide via click chemistry (Salic and Mitchison 2008). Medaka was exposed to a mixture of 5-FU and E-dU and then stained with anti-BrdU (for 5-FU) and Alexa 488–conjugated azide (for E-dU). Figure 2 shows that the tissue distribution patterns of 5-FU and E-dU were completely different. Significantly, while medaka eye (Figures 3A–3D), gut (Figures 3E–3H), and brain (Figures 3I–3L) all displayed widespread 5-FU incorporation, only gut cells showed a significant labeling by E-dU, consistent with the restriction of DNA synthesis activities to a small population of proliferating cells. In gut, some strongly 5-FU–labeled cells were E-dU negative, whereas some E-dU–positive cells were not labeled by 5-FU. This suggests that the incorporation of these two nucleoside analogs was highly dependent on the activities of individual cells, rather than on the general biodistribution and uptake of nucleosides. In brain, there were a small number of E-dU–positive (proliferating) cells. A higher magnification (Figures 3M–3P) shows 5-FU–labeled and E-dU–labeled distinct, non-overlapping sets of nuclear foci, reminiscent of the non-colocalizing transcription factories and DNA replication sites in mammalian nuclei (Wei et al. 1998). To our knowledge, this is the first report of colabeling of transcription and replication sites at the organismal level.

Whole-mount Labeling of Transcription Factories in Medaka

Our method of labeling nascent RNA in medaka provides a valuable system for studying transcription factories at the tissue level. To illustrate this point, we focused our investigation on two important organs: skin and eye. Because it was difficult to examine the body surface on sections, we explored the use of whole-mount immunofluorescence in revealing transcription sites in skin cells. Figure 4A shows the whole-mount labeling of 5-FU incorporation in medaka. All nuclei of skin and fin cells were brightly stained, and could be examined in great detail (Figures 4B–4D). In some cells, 5-FU was incorporated in large nuclear structures resembling the nucleoli, consistent with the abundant rRNA synthesis in the nucleolus (Lam et al. 2005). In addition, there were hundreds of small nucleoplasmic foci that resembled mammalian transcription factories. To show that the antibodies could penetrate deep into the fish, we examined the specimen using multiphoton confocal microscopy (Figures 4E–4H). Interestingly, levels of 5-FU incorporation appeared to be correlated with the cell morphology, inasmuch as cells with thin, elongated nuclei invariably displayed the highest 5-FU labeling, and cells with round, condensed nuclei had undetectable 5-FU incorporation.

Transcription Factories on Medaka Retina

We consistently observed an interesting distribution pattern of 5-FU labeling on medaka retina (Figures 1D, 1L, 1M, and 2A), in which 5-FU incorporation was more brightly stained in the inner and outer layers than in the middle layer, suggesting a remarkable correlation of levels of RNA synthesis with the retinal architecture. The structure of teleost retina has been extensively studied. Medaka retina is composed of layers of cells, organized from the basal layers toward the lens (Kitambi and Malicki 2008). The most-basal layer is the retinal pigment epithelium, followed by three layers of neurons: photoreceptor cell layer, inner nuclear layer, and ganglion cell layer. To investigate the identities of cell layers that incorporated 5-FU strongly, we labeled serial transverse sections of a medaka eye with anti-BrdU and with anti-ZPR1, a marker of the outer nuclear layer of medaka (Chen et al. 2009). The coincidence of strong 5-FU incorporation with anti-ZPR1 immunostaining (Figures 5A and 5D) indicated that the photoreceptor cells in the outer cell layer were highly transcriptionally active, whereas the bipolar, horizontal, and amacrine cells of the inner cell layer appeared to be transcriptionally quiescent. Interestingly, the outermost retinal layer, possibly the ganglion cell layer, was brightly 5-FU positive. To further visualize the topographical distribution of transcriptionally active cells in the retina, we investigated the pattern of 5-FU incorporation in tangential sections of the medaka retina (Figures 5G–5J). Interestingly, the transcriptionally active cells in both outer and inner nuclear layers were arranged in a mosaic pattern. This suggests that the transcription rates in retinal cells are tightly regulated and may be associated with the growth and functioning of these cells.

Discussion

This study establishes the marine medaka fish as a model organism for high-resolution characterization of transcription factories at the organismal level. Whole-body labeling of nascent RNA in rodents has been reported. In these studies, rodents injected with tagged nucleoside were killed and dissected, and nascent RNA in various tissues was revealed by autoradiography (Uddin et al. 1984), immunocytochemistry (Casafont et al. 2006), or click chemistry (Jao and Salic 2008). These few studies all show that the overall rate of RNA synthesis appeared to be highly cell type specific, in a manner comparable to the tissue-specific pattern of DNA replication, thus suggesting a previously overlooked aspect of cellular activity. However, the use of small mammals in metabolic labeling experiments is often complicated by the complex pharmacokinetics of the injected precursor. We believe an alternative experimental model is required for the study of the tissue distribution of transcriptional activities.

The marine medaka is an excellent model for cell and developmental biology. The juvenile medaka fish used in this study are small (less than 5 mm in length) and can be raised in large quantities. Of more importance, these fish are relatively permeable to exogenous chemicals such as fluorescent probes (Hardman et al. 2008). We utilized this property and showed here that when put in ambient water, 5-FU could permeate throughout the body of medaka. The direct mode of uptake, combined with the small body size of medaka, allows for relatively simple in vivo distribution kinetics of nucleosides. Furthermore, we demonstrated here that transcription sites of the entire medaka fish could be visualized by whole-mount immunofluorescence. With the aid of multiphoton confocal microscopy, the cell type–specific variation of total transcription rates in the entire medaka body can be imaged at subcellular resolution and reconstructed three-dimensionally. Taken together, our results indicate that the marine medaka is a more-favorable model organism, compared with rodents, for the study of tissue-specific regulation of RNA synthesis.

We demonstrated here that 5-FU labeling could be conveniently combined with E-dU in the simultaneous labeling of DNA replication sites and RNA transcription sites at high resolution. It would be interesting to investigate the spatial organization of these two essential processes in different cell types, especially during early embryonic development. In this context, the ex-utero embryonic development of medaka offers an ideal experimental system.

In this study, we demonstrated the intriguing pattern of transcriptional activity in fish retina. The inner nuclear layer appeared to be transcriptionally quiescent, whereas the outer nuclear and ganglion cell layers were among the tissues with the strongest 5-FU incorporation in the medaka body. It is not clear how transcription rates are correlated with the physiology of these cell types. The architecture, development, and regeneration of retina in teleost fish have been intensively studied (Cameron 2000), and a large number of genetic tools are available in medaka (Loosli et al. 2004). It will be important to study the effects of mutations associated with retinal malformation on transcription levels, and the possible changes in transcription factories during retinal development and repair in medaka.