|Home | About | Journals | Submit | Contact Us | Français|
Crx is a member of the Otx family of homeobox genes with expression restricted to vertebrate retinal photoreceptor and bipolar cells as well as the pinealocytes of the pineal organ. To facilitate the visualization of Crx-expressing cells, we generated transgenic mice expressing several reporters under the control of the Crx regulatory sequences present within a bacterial artificial chromosome (BAC). These mice expand the transgenic mouse collection, which uses photoreceptor regulatory elements for reporter gene expression by providing a broader repertoire of reporter genes. In addition, since Crx is expressed very soon after a cell fated to be a photoreceptor cell becomes postmitotic, they provide a means for early identification of immature photoreceptor cells.
Reporter lines that facilitate the histochemical and/or live detection of specific cell types are generally useful for studies of development and function. Photoreceptor cells are a critical cell type for vision and have been well studied by a variety of methods. Strains of mice that express reporter genes specifically in photoreceptors have been previously constructed. These include lines that employ the Rbp3 (IRBP) promoter to express β-galactosidase in rod and cone photoreceptors (Yokoyama et al., 1992), as well as the cGMP-phosphdiesterase beta (Ogueta et al., 2000), Rhodopsin Kinase (Young et al., 2003), and Qrx promoters (Wang et al., 2004). The promoters of the photopigment genes, namely Rhodopsin (Zack et al., 1991; Li et al., 2005) and Cone Opsins (Chiu and Nathans, 1994; Fei and Hughes, 2001) have been used to construct reporter strains for rod or cone photoreceptors, respectively. In addition, the Nrl promoter was recently used to drive GFP and thus mark rod photoreceptors, specifically (Akimoto et al., 2006). Many photoreceptor genes, including the photopigment genes, are not expressed in the early stages of differentiation of photoreceptors (Blackshaw et al., 2004), which is when Nrl-GFP (Akimoto et al., 2006) and Crx expression begins.
Crx is a member of the Otx family of homeobox genes (Chen et al., 1997; Furukawa et al., 1997; Plouhinec et al., 2003). Its expression is quite restricted, having been previously observed only in vertebrate retinal photoreceptor cells, retinal bipolar cells, and the pinealocytes of the pineal organ. Crx mutations in humans and mice have shown it to be necessary for the differentiation and function of rod and cone photoreceptors (Freund et al., 1997; Swain et al., 1997; Furukawa et al., 1999; Rivolta et al., 2001), as well as the regulation of melatonin synthetic enzymes of the pineal organ (Furukawa et al., 1999). We generated transgenic mice expressing several reporters under the control of the Crx regulatory sequences present within a bacterial artificial chromosome (BAC) encompassing the Crx locus. These mice allow for the identification of Crx-expressing cells using one of three reporter genes with unique properties for histochemical, immunochemical, or fluorescence detection studies.
A 121 kilobase (Kb) BAC from Incyte Genetics was used to make two reporter constructs. Sequencing (see Materials and Methods) of the ends of the BAC showed that it contained 21 Kb 5’ of the start site of the Crx gene (Furukawa et al., 2002), and 86 Kb 3’ of the most often used retinal polyadenylation site (Furukawa et al., 2002; Hodges et al., 2002). One construct, CRXLacZAP, encodes nuclear β-galactosidase (nβGal), with its translation initiating at the Crx ATG, and includes human placental alkaline phosphatase (PLAP) under the control of an internal ribosomal entry site (IRES) from ECMV (encephalomyocarditis virus from rabbit). The other construct, CRXGFPAP, encodes green fluorescent protein (GFP), with its translation initiating at the Crx ATG, followed as well by IRES-PLAP (Figure 1). Both sets of bicistronic transgenes replaced exon 2 using homologous recombination (Muyrers et al., 1999; Lee et al., 2001). Purified BACs containing the targeting constructs were injected into the pronuclei of fertilized SJLBL/6 blastocysts and founders were obtained. One founder was obtained for CRXLacZAP and five for CRXGFPAP. Founders for both constructs were backcrossed to C57BL/6 animals for maintenance as lines.
To test for nβGal and PLAP activity in the CRXLacZAP line, embryos were harvested each day from E12.5-E16.5 and stained histochemically using a X-gal stain to detect nβGal activity and an AP stain to detect PLAP activity. All embryonic time points of the CRXLacZAP line exhibited X-gal and AP staining in the predicted pattern of cells fated to become photoreceptors. These cells are located at the scleral edge of the retina, the location of the future outer nuclear layer (ONL), where mature photoreceptor cell bodies will reside. Staining was robust, with visible precipitate appearing within one hour of staining. To compare endogenous Crx expression to that of the reporters, in situ hybridization for (RNA) Crx was performed on non-transgenic, wild-type (WT) animals and compared to X-gal and AP staining of transgenic animals. At the embryonic time points tested (E12.5, E13.5 and E15.5 only shown), nβGal and PLAP (2E, 2F, 2H & 2I) activity closely followed the expected in situ hybridization pattern of endogenous Crx RNA in the position of cells predicted to be newborn and maturing photoreceptor cells (2B & 2C). However, one unexpected aspect of the staining with the reporters was noted at E12.5 in that X-gal staining was observed surrounding the optic nerve head (ONH) (Figure 2D) and AP staining was present in the ONH (2G). Both of these staining patterns do not correspond to the predicted location of photoreceptor cells. Therefore, it was possible that this staining was a result of ectopic activity of the reporters, perhaps due to the use of a BAC transgenic. In situ hybridization for Crx was thus carried out on sections including the ONH (Figure 2A), and signal was indeed observed for Crx RNA in the area of the ONH, as shown by the reporters. The reporters thus faithfully recapitulated endogenous Crx expression and revealed a hitherto unrecognized expression in what is likely to be non-photoreceptor cells (see Discussion).
The expression of nβGal and PLAP in early postnatal animals was analyzed by both wholemount staining and section staining. At postnatal day 6 (P6), strong staining was present in the developing photoreceptor layer of the retina with visible precipitate appearing within an hour of staining (Figure 2J & 2K). Adult retinae and adult brain also were assayed for reporter expression. Expression of both nβGal and PLAP were robust in the retinae, with staining evident within an hour of addition of staining solution. X-gal staining of the adult transgenic retinae was strong in inner segments (IS), ONL, and the inner nuclear layer (INL), where the bipolar cell bodies are located (Figure 3A). Some weak X-gal staining was present in the ganglion cell layer, which could be ectopic staining of the transgene or reflect weak Crx expression in ganglion cells (see Discussion). AP histochemical staining followed the same pattern, with additional AP staining in the outer segments (OS) and in the inner plexiform layer (IPL) which is likely to be from bipolar processes (Figure 3B). However, out of the dozen retinae stained from the first two generations of this founder, two had a mosaic pattern: one adult, shown in Figure 3C-D, and one P6 pup (Figure 2L). The mosaic pattern was evident in both the photoreceptor layer and in the INL. More adult retinae, but not all, taken from later generations (generations six and eight) exhibited the mosaic pattern of X-gal and AP staining as described above. WT animals were negative for nβGal and PLAP activity when stained for these reporters (3E,F).
In adult transgenic brains of early generations, nβGal and PLAP expression was observed in the pineal gland, where Crx is known to be present (Furukawa et al., 1997), while staining was absent from non-transgenic animals (data not shown). However, brains taken from later generations of this line no longer stained for X-gal or AP in the pineal gland.
Expression of GFP and PLAP were assayed in the CRXGFPAP transgenic line at embryonic and adult stages. GFP fluorescence could be observed in wholemount E13.5 embryos, both in the eye and in the developing tectum (Figure 4A & 4B). Confocal imaging of eye sections revealed that the majority of GFP-positive cells could be observed along the scleral edge of the retina where photoreceptors positive for Crx RNA are found (Figure 4C). Like Crx, expression of GFP in isolated cells was observed outside of this layer in the retina. Expression could also be observed in cells adjacent to the ONH, again recapitulating Crx expression. Expression of GFP in the retinal pigmented epithelium (RPE) was also observed, especially near the ONH where pigmentation was somewhat incomplete. Crx mRNA expression has not been observed in this tissue before, so whether this is ectopic expression of the transgene, visualization of weak Crx expression with the sensitive GFP, or some other phenomena, we cannot conclude at this time. A novel site of Crx expression was also observed in the tectum. Sections through the tectum revealed CrxGFPAP expression in a subset of cells in this tissue that closely resembled the endogenous expression of Crx RNA (Figure 4E & 4F). At this embryonic timepoint, the eye and tectum expression was easily visualized using GFP fluorescence. Immunofluoresence detection of GFP allowed for additional identification of a small number of sites of expression of the transgene that could not be seen with simple GFP fluorescence (data not shown). In the adult eye, strong expression of GFP could be observed in both wholemount preparations of the retina and in the ONL, IS and part of the INL on retinal sections (Figure 4G-I). For PLAP, expression overlapped with GFP with additional PLAP expression present in the OS (Figure 4J). In adult brain, PLAP was strongly expressed in the pineal gland and persisted in later generations (5A-B). No AP stain was detected in WT animals (5C-D).
Expression of Crx RNA can be observed at E12.5 in mouse cells that appear to be fated to be rod and cone photoreceptors. Its expression peaks at approximately P6 and is maintained at a lower level in adult photoreceptors. It is also expressed at relatively low level in bipolar cells (Liu et al., 2001). In other tissues, expression has been noted only in the pineal gland, which contains cells similar to photoreceptor cells (Li et al., 1998). The transgenic reporter strains reported here recapitulated all of these aspects of Crx expression. In early generations of the transgenic lines, expression was robust in all stages of photoreceptor development and in the pineal gland. However, in later generations, the pineal expression appears to have been lost in the CRXLacZAP line but persisted in the CRXGFPAP line. In bipolar cells, the expression in all generations is quite weak, but detectable, as is expression of endogenous Crx RNA. Two previously unappreciated domains of Crx expression were noted in these lines. Near and in the optic nerve there was heavier expression of the reporters than was previously detected by in situ hybridization for Crx. When in situ hybridizations were performed to scrutinize expression in this area, expression of endogenous Crx RNA could be seen. The more robust expression revealed by the reporters likely is due to the greater sensitivity of the reporter gene activity. Recently we have performed comprehensive RNA profiling of single cells from the developing retina (Trimarchi et al., 2008). One of these cells was a newborn ganglion cell, and it too expressed Crx, at a low level. It is thus possible that the cell type that expresses low levels of Crx near the ONH or the ganglion cell layer staining detected in Figure 3A are ganglion cells. Another strong possibility based on morphology and location is that the ONH cells are optic disc astrocyte precursor cells that have been described both morphologically and molecularly (Nornes et al., 1990; Petros et al., 2006). A novel expression pattern for Crx was also detected in the developing tectum from both the transgenic reporter and for endogenous Crx RNA.
The use of three different genes for reporter activity should make these strains useful for several applications. GFP is obviously useful for applications in which live visualization is desired, and it is also useful for double staining using other fluorescent reagents (e.g. to detect a protein epitope that might be expressed in photoreceptor cells). PLAP offers great sensitivity. In the course of creating many different types of reporters for use in viral vectors, we have found that PLAP is the most sensitive reporter (Fields-Berry et al., 1992). This may be due to the fact that the enzyme must be quite stable, withstanding heating to 65°C, which is done to reduce endogenous AP activity. This stability is likely why it continues to cleave substrates for more than one day in the histochemical reaction conditions. PLAP is a gpi-linked protein (Kam et al., 1985), which allows for labeling of the plasma membrane, revealing details such as the axonal endings of bipolar cells. It is also valuable if one wishes to examine, for example, the morphological transformations that photoreceptor cells undergo as they differentiate. The nβGal reporter is of use when one wishes to quantify cells, as nuclear stains facilitate counting, and for localization of photoreceptor cell bodies. It is also useful for co-localization studies, e.g. when one wishes to double stain the nucleus.
These lines also allow for the conclusion that the majority, or perhaps all, of the regulatory information for proper expression of Crx is localized to the BAC that was used here. The BAC contains 21Kb 5’ and 86 Kb 3’ of the Crx gene. Previously, it was shown that as little as 2 Kb 5’ of Crx , along with the first intron, could confer photoreceptor specificity (Furukawa et al., 2002). However, even up to 12 Kb 5’ of the Crx gene did not direct high level synthesis of βGal as was found with the BAC used here. The high level of expression afforded by the BAC enables detection of GFP, and thus prospective isolation of newborn or mature photoreceptors. This BAC also provides an opportunity to dissect the regulatory elements required for high level and specific expression in developing and mature photoreceptor cells.
These mice have been deposited at Jackson Labs and are available to all.
A 121 Kb BAC from Incyte Genetics was characterized by sequencing of the ends, with the 5’ and 3’ ends designated as such relative to the 5’ and 3’ ends of the Crx gene.
5’ end on chr7:7,521,580-7,522,226:
3' end on chr7:7,642,839-7,642,922:
For both the CRXLacZAP and CRXGFPAP targeting vectors, IRES-PLAP was excised from pLIA (Bao and Cepko, 1997) with NotI and SalI and cloned into NotI and SalI sites of pBluescriptIISK to create pBS-PLAP. The EcoRV site was reintroduced into pBS-PLAP with linkers into the SnaBI site of pBS-PLAP to facilitate cloning of the nlacZ PCR product. This construct was designated 5’ homology arms and cloning sites (sequences below) were added by PCR to both nuclear lacZ (nlacZ) and GFP genes and cloned upstream of IRES-PLAP in pBS-PLAP and pBS-PLAPΔ. For CRXLacZAP, the nlacZ PCR product was digested with NotI and EcoRV and inserted into the NotI and EcoRV sites of This construct was designated pBS-CRX(Z). For CRXGFPAP, the GFP PCR product was digested with NotI and SnaBI and inserted into the NotI and SnaBI site of pBS-PLAP to create pBS-CRX(G). For antibiotic selection, a FRT flanked kanamycin/neomycin cassette containing both 3’ homology arms and restriction sites for cloning were generated by PCR and cloned into the XhoI sites of both pBS-CRX(Z) and pBS-CRX(G). These final constructs, both containing NotI sites for excision of the targeting construct, were used for recombination and named pCRXLacZAP and pCRXGFPAP.
pBS-CRXLacZAP and pBS-CRXGFPAP targeting cassettes were liberated with NotI from BluescriptIISK. DNA was gel purified and relevant fragments excised and extracted in preparation for recombination into the CRX BAC clone. Each targeting cassette was recombined into the BAC clone using an E-T homologous BAC recombination method (Muyrers et al., 1999), with the modification of Lee et al. (Lee et al., 2001). Several homologous recombined clones for each targeting construct were obtained and one was chosen for injection. Prior to injection, BAC clones for CRXLacZAP and CRXGFPAP were digested with NotI to liberate the BAC DNA from the backbone. Digested BAC DNA was run on a 0.6% gel for purification and then extracted by electrophoresis. DNA was then collected in a small volume of running buffer, precipitated and resuspended in a small volume for injection. CRXLacZAP and CRXGFPAP BAC DNA were then injected into the pronuclei of SJLBL/6 blastocysts and founders were obtained.
One founder was obtained for CRXLacZAP and was positive for the transgene by PCR and Southern blot. Thirty-six founders were born for CRXGFPAP and five were positive for the transgene by PCR and Southern blot. Out of the five founders, three produced offspring. Of the three, one was used for the analysis presented in this study.
5’ CRXLacZAP and CRXGFPAP homology arm targeted to the second exon:
3’ CRX homology arm targeted to the second exon:
5’ Primer for Genotyping CRXLacZAP mice:
3’ Primer for Genotyping CRXLacZAP mice:
5’ Primer for Genotyping CRXGFPAP mice:
3’ Primer for Genotyping CRXGFPAP mice:
PCR for CRXLacZAP Conditions:
PCR for CRXGFPAP Conditions:
Embryos were fixed for one hour in 4% paraformaldehyde (PFA), sectioned at 20 microns and processed for X-gal staining. For adult retinal tissue, most retinae were fixed in 4% PFA. For increased sensitivity of the X-gal and AP stain, some retinae were fixed in 0.5% glutaraldehyde (Panels A and B in Figure 3) ( http://genetics.med.harvard.edu/~cepko and (Ausubel et al., 1997)). Blue X-gal precipitate was noticeable after 15 minutes for retinal sections. For AP staining, embryos were treated and sectioned as above http://genetics.med.harvard.edu/~cepko and (Ausubel et al., 1997)).
Brains were removed from adult animals (older than P35) and flash frozen in cooled isopentane (-70 °C). Cryosections were collected at 20 microns and dried overnight before storage at –20 °C or staining. Staining for nβGal and PLAP was done as above and stained overnight. Tissues were then counterstained with Eosin.
Twenty micron transverse sections were generated by cryosectioning. Dapi was used to visualize nuclei and imaging was performed on a Leica SP2 inverted confocal microscope. 5.7 micron stacks were collected and average projections were generated with ImageJ (version 1.40g).
Crx probes were generated using clone BE955622 and hybridizations were performed using methods previously detailed (Trimarchi et al., 1997).
We thank Sheldon Rowan for providing panels C and D in Figure 3, J.P. Antoine Grande for performing the in situ hybridizations in panels B and C in Figure 2, Andrew Abney for pronuclear injections and Amy Shaw for animal husbandry.
Grant Sponors and Grant Numbers National Institutes of Health - F32 EY018287-01 Howard Hughes Medical Institute