Insertion of a Pax6 Consensus Binding Site into the αA-Crystallin Promoter Acts as a Lens Epithelial Cell Enhancer in Transgenic Mice

Abstract

purpose. Although the murine αA-crystallin promoter is the most commonly used promoter for achieving transgene expression in the developing lens, this promoter directs transgene expression efficiently only in lens fiber cells. The purpose of the present study was to generate promoters capable of directing transgene expression to the entire lens but not to the corneal epithelium.

methods. Transgenic mice were generated with fragments of the murine αA- and αB-crystallin promoters, as well as with an αA-crystallin promoter engineered with the insertion of a Pax6 consensus binding site driving either human growth hormone (hGH) or Cre recombinase genes. hGH expression was evaluated by in situ hybridization and immunohistochemistry. Cre expression was revealed by x-gal staining after crossing Cre transgenic mice with a Cre reporter strain.

results. Within the lens, the −214/+38 αB-crystallin promoter fragment directed transgene expression in the lens epithelium, but not in fiber cells. The native −282/+43 αA-crystallin promoter drove transgene expression in the lens fiber cells of several independent lines of transgenic mice, but none of these mice demonstrated significant transgene expression in the lens epithelium. In contrast, the insertion of a 32-bp sequence containing a Pax6 consensus binding site into the −282/+43 αA-crystallin promoter reproducibly led to transgene expression in the lens epithelium as well as the lens fiber cells.

conclusions. The inclusion of a Pax6 consensus binding site within the −282/+43 αA-crystallin promoter enhances the ability of this promoter to drive transgene expression in the lens epithelium.

The lens consists of a proliferative epithelium that lines the anterior surface of the structure, covering the apical surface of a core of postmitotic terminally differentiated fiber cells. These two types of cells represent a single lineage, in that lens epithelial cells migrate toward the equatorial region of the lens where they exit the cell cycle and differentiate into fiber cells. Genetic engineering has emerged as a powerful tool for the investigation of the molecular basis of lens fiber differentiation. The mouse lens was one of the first tissues in which tissue-specific transgene expression was demonstrated ¹ and was the first tissue in which Cre-mediated recombination was used to catalyze tissue-specific genomic engineering. ² Most of the promoters used for transgene expression in the lens are derivations of crystallin promoters, and with the exception of the αB-crystallin promoter, ^{3 4 5} they are capable of driving high-level transgene expression only in the lens fiber cells. ^{6 7 8 9 10 11 12 13} To investigate mechanisms of fiber differentiation, however, it is desirable to manipulate gene expression in the undifferentiated lens epithelium. 

Crystallins are major water-soluble cytoplasmic proteins in the lens. The α-, β-, and γ-crystallins, found in all mammalian lenses, are expressed in either a lens-specific or a lens-preferred manner. The α-crystallins consist of αA- and αB-crystallin. The lens is the major site of αA-crystallin expression, where it is detected in both epithelial and fiber cells, but a dramatic increase in the expression of αA-crystallin occurs as the lens epithelial cells differentiate into lens fibers. ¹⁴ Only trace amounts of αA-crystallin have been detected in a small number of non–lens tissues. ^{15 16} Multiple cis-acting transcriptional regulatory elements residing in the 5′-flanking region of the murine αA-crystallin gene contribute to its lens-preferred expression. ^{17 18 19 20 21 22}  

In contrast, murine αB-crystallin is expressed in the lens, retinal pigment epithelium (RPE), heart, skeletal muscle, and brain. ^{3 23 24 25 26 27 28} A short fragment of the αB-crystallin promoter (−164/+44), lacking upstream elements shown to be important for non–lens expression of αB-crystallin, ^{25 26 27} was shown to confer lens specificity in transgenic mice. ²⁹ However, a later report demonstrated that this promoter is also active in the postnatal corneal epithelium. ⁵ Two regulatory elements within this region bind Pax6 and mediate transcriptional activation of the αB-crystallin gene in the lens epithelium. ⁴ Unlike αA-crystallin, αB-crystallin has more pronounced expression in the lens epithelium than in the fibers. ¹⁴  

In the embryonic and adult lens, Pax6 mRNA is detected mainly in the epithelium. ^{30 31} Although Pax6 protein is detectable in newly differentiated lens fibers, the protein is gradually lost as fiber differentiation proceeds. ^{30 31} Pax6 plays many critical roles in early vertebrate eye development, and lens formation does not occur in the absence of Pax6. ^{32 33} Pax6 plays a role in regulating the spatial and temporal expression of αA-, αB-, γE-, and γF-crystallins in the mouse; αA-, βB1- and δ1-crystallins in chick, and ζ-crystallin in the guinea pig. ^{4 17 31 34 35 36 37} In addition, Pax6 has been shown to act upstream of transcription factors that are essential in lens development including Sox2, Six3, L-Maf, and Prox1. ^{32 38 39}  

An element from the Pax6 P0 promoter, known as the ectoderm enhancer, has been shown to direct transgene expression in the lens placode and placode-derived structures, such as the epithelium of the lens, cornea, conjunctiva, and lacrimal gland. ^{40 41} Transgene expression with this enhancer is not maintained in lens fiber cells. Our purpose was to find or engineer promoters capable of driving transgene expression in the entire lens, including the lens epithelium, but without the inclusion of other ocular epithelia. We chose to use two different genes to evaluate the promoter expression pattern to control for possible influences of coding sequence on promoter activity. The first of these, human growth hormone (hGH), is not normally expressed in the eye, but can easily be detected by in situ hybridization or immunohistochemistry. The second gene was Cre recombinase. Cre is a P1 bacteriophage-derived DNA recombinase with a specific 34-bp recognition sequence called a loxP site. Cre recombinase was used to test these promoters with the thought that the resultant mice could not only provide more in vivo evidence about the effect of the Pax6 consensus binding site on the αA-crystallin promoter activity, but also could serve as tools to create conditional inactivation of genes in a tissue-specific manner. 

Within the eye, transgenic constructs containing the murine −214/+38 αB-crystallin promoter most often drove transgene expression in the lens epithelium and RPE, whereas constructs containing the native murine −282/+43 αA-crystallin promoter most often drove transgene expression only in fiber cells of the lens. Insertion of a DNA sequence that binds Pax6 ³⁴ into the αA-crystallin promoter enhanced the expression of two different reporter genes in the lens epithelium while maintaining fiber cell expression. 

Methods

Construction of Transgenic Vectors

All primer sequences and detailed methods are available in the online appendix. A novel BglII site was inserted into the CPV2 vector ⁴² at −86 relative to the murine αA-crystallin transcription start site by overlapping PCR, creating the intermediate CPV7 vector. A consensus Pax6 binding site, ³⁴ created by annealing two partially complementary oligonucleotides with 5′ overhangs compatible with the BglII restriction site, was inserted into the BglII site of CPV7 to generate CPV14 and CPV15 vectors, differing only by the orientation of the single Pax6 binding site. A genomic fragment containing the hGH gene from pOGH (obtained from Francesco DeMayo, Baylor College of Medicine, Houston, TX ⁴³ ) was subcloned into CPV2, CPV14, and CPV15, creating CPV2/hGH, CPV14/hGH, and CPV15/hGH, respectively. 

To create the αB1/hGH construct, genomic sequence from −214 to +38 of the mouse αB-crystallin gene (relative to the major lens transcription start site, GenBank accession no. M73741 ⁴⁴ ; http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD) was amplified by PCR and inserted into CPV2, replacing the murine αA-crystallin promoter and creating the vector αB1. The promoter of αB1 was ligated into the vector backbone of CPV2/hGH, replacing the αA-crystallin promoter to create αB1/hGH. 

The αA/αB4/hGH construct was made through several steps. First, the 3′ untranslated region and polyadenylation signal of the murine αB-crystallin gene was amplified by PCR and ligated into the large fragment of the αB1 vector, replacing the SV40-derived intron and polyadenylation sequence of αB1. This intermediate plasmid was called αB2. The first intron of the murine αB-crystallin gene was amplified by PCR and inserted into the αB1 vector, creating the intermediate vector αB3. A three-way ligation with the αB-crystallin promoter and first intron from αB3, the hGH gene from pOGH, ⁴³ and the αB-crystallin 3′ untranslated region–polyadenylation sequences and plasmid backbone of αB2 created αB4/hGH. To generate the composite αA/αB promoter, CPV7 (described earlier) was digested with BglII and BamHI to remove the −86/+43 region of the αA-crystallin gene. The PCR-amplified murine αB-crystallin promoter (described earlier) was digested with BglII and ligated into the BglII/BamHI-cut CPV7, creating the intermediate αA/αB composite promoter vector. The αA/αB composite promoter was then used to replace the αB-crystallin promoter of αB4/hGH, completing the αA/αB4/hGH construct. 

The Cre coding sequence and intron–polyadenylation sequences contained in the murine metallothionein gene were excised from pBS216 (obtained from Brian Sauer, Stowers Institute, Kansas City, MO ² ) and ligated into HindIII/SalI cut CPV14 and CPV2, creating CPV14/Cre and CPV2/Cre, respectively. 

Generation of Transgenic Mice

All animals were treated in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Microinjection fragments were isolated from CPV2/hGH, CPV14/hGH, CPV15/hGH, αB1/hGH, TYBS (a tyrosinase minigene), ⁴⁵ αA/αB4/hGH, CPV14/Cre, and CPV2/Cre following digestion with appropriate restriction enzymes and gel purification with a gel extraction kit (QiaexII; Qiagen, Hilden, Germany). All microinjection constructs were injected into pronuclear stage FVB/N mouse embryos as described. ⁴⁶ All hGH transgenic constructs were injected independently, but Cre constructs were either injected independently or coinjected with the TYBS cassette. ⁴⁵  

Histology, Immunohistochemistry, and In Situ Hybridization

Eyes or embryos were collected and fixed in 4% paraformaldehyde, processed, and embedded in paraffin. Embedded samples were sectioned at 5 μm. Immunohistochemical staining for hGH was performed as described previously. ⁴³ An hGH-specific riboprobe vector, CB4, was created by ligating the 3′ untranslated region of the human growth hormone gene from the pSW2 ⁴⁷ vector (obtained from Lewis T. Williams, University of California, San Francisco) into pBluescript II KS(−) (Stratagene, La Jolla, CA). Antisense [³⁵S]-UTP–labeled probes were synthesized using a HindIII-digested CB4 template and T7 RNA polymerase (Promega, Madison, WI). In situ hybridization of tissue sections was performed as described previously. ⁴²  

X-Gal Staining for β-Galactosidase Activity

To evaluate the expression pattern of the Cre transgene, Cre transgenic mouse lines were crossed with a homozygous ROSA26 reporter line (B6;129S-Gtrosa26 ^tm1Sor ). ⁴⁸ Mouse embryos, neonatal eyes and lens epithelium from mice at weaning age were collected and analyzed for β-galactosidase activity by x-gal staining, as described. ⁴²  

Results

Characterization of Engineered α-Crystallin Promoters

We produced several transgenic constructs in which an hGH reporter gene was placed under the regulatory control of modified versions of the murine αA-crystallin promoter or a short fragment (−214/+38) of the murine αB-crystallin promoter (Table 1) . The activity of these promoters was initially tested by using a transient transgenic approach in which transgenic founder embryos were collected at 15.5 days after coitus (embryonic day [E]15.5). At least two transgenic founder embryos were generated for each construct and evaluated for hGH expression by both in situ hybridization and immunohistochemistry. 

CPV2/hGH consisted of the hGH reporter gene downstream of the (−282/+43) murine αA-crystallin promoter (Fig. 1A , Table 1 ). The αB1/hGH construct was identical to CPV2/hGH except that the αA-crystallin promoter was replaced with a short fragment (−214/+38) of the murine αB-crystallin promoter (Fig. 1A , Table 1 ). A composite promoter including the −282/−86 region of the murine αA-crystallin promoter followed by the −214/+38 promoter fragment and the first intron of murine αB-crystallin gene was also created. The hGH reporter gene and 3′ untranslated region of αB-crystallin were placed downstream of this chimeric promoter, creating αA/αB4/hGH (Fig. 1A , Table 1 ). Prototypical expression patterns for these founders are shown in Figures 1B 1C 1D 1E 1F 1G 1H 1I 1J 1K 1L 1M . In CPV2/hGH transgenic animals, transgene-specific mRNA and hGH protein expression was detected in the lens fiber cells but not in the lens epithelium (Figs. 1B 1C 1D 1E) , typical of the activity of the −282/+43 fragment within the αA-crystallin promoter. ^{42 49 50 51} Of three founder embryos carrying the αB1/hGH transgene, only one (33%) showed evidence of reporter gene expression. The −214/+38 region of the αB-crystallin promoter directed hGH expression in this founder embryo to the lens epithelium and RPE with very little detectable expression in lens fiber cells (Figs. 1F 1G 1H 1I) . Three stable transgenic lines were generated with the αB1/hGH construct (OVE1046, OVE1047, and OVE1048). Of these lines, two (OVE1046 and OVE1048) demonstrated transgene expression in the lens epithelial cells, RPE, and developing ciliary body in E15.5 embryos. One line (OVE1046) expressed the reporter gene weakly in the lens fiber cells (data not shown). Overall, three (50%) of the six transgenic founders or lines made with the αB1/hGH construct expressed the transgene in the lens. In an attempt to increase the level of lens transgene expression in the αB1/hGH founders, we created a chimeric promoter consisting of the −282/−86 fragment of the mouse αA-crystallin promoter fused to the −214/+38 mouse αB1-promoter, intron 1 of αB-crystallin, and the 3′ untranslated sequence. Of the five αA/αB4/hGH transgenic founder embryos, reporter gene expression was found in three (60%). In these founders, hGH was detected in the lens epithelium and RPE. In one (20%) founder, patchy expression in the corneal epithelium was detected by in situ hybridization, but not by immunohistochemistry (Figs. 1J 1K 1L 1M) . None of the αA/αB4/hGH founder embryos expressed the transgene in the lens fiber cells. Thus, despite its activity in the lens epithelium, the αA/αB4 chimeric promoter failed to activate transgene expression in the lens fibers, indicating the −282/−86 αA-crystallin promoter fragment was insufficient to activate transgene expression in the lens fibers when coupled with the −214/+38 αB-crystallin promoter. 

Effect of Insertion of the Pax6 Consensus Binding Site into the αA-crystallin Promoter on Transgene Expression in the Lens Epithelium

The differential expression patterns of αA- and αB-crystallins within the lens and the observation that Pax6 directly regulates expression of αB-crystallin in the lens epithelium prompted us to hypothesize that we might be able to enhance the activity of the murine αA-crystallin promoter in the lens epithelium by incorporating additional Pax6-binding regulatory elements. To test this hypothesis, we made a small modification in CPV2 by inserting a 32-bp sequence containing a Pax6 consensus binding site in either orientation, creating CPV14 and CPV15, respectively. The insertion site is located at −86 (relative to the transcription start site) of the αA-crystallin gene, a region devoid of elements known to be important for αA-crystallin transcriptional regulation. The hGH gene was then subcloned into both of these vectors, creating CPV14/hGH and CPV15/hGH (Fig. 2A , Table 1 ). 

Transgenic mice made with either CPV14/hGH or CPV15/hGH constructs expressed hGH mRNA and protein specifically in both the lens epithelium and fibers (Figs. 2B 2C 2D 2E 2F 2G 2H 2I) . Because both CPV14/hGH and CPV15/hGH supported transgene expression in the lens epithelium, the inclusion of the Pax6 consensus binding sequence in the −282/+43 region of αA-crystallin promoter enhanced lens epithelial transgene expression in an orientation-independent manner. In these transgenic founder embryos, no ocular expression of the hGH reporter gene was detected outside the lens. 

Generation of CPV2/Cre and CPV14/Cre Transgenic Mice

Because hGH expression driven by both the CPV14 and CPV15 promoters was specific to the lens and detected in both the lens epithelium and fiber cells in transgenic founder embryos, we decided to test the activity of the CPV14 promoter more thoroughly by using a different reporter gene. To this end, we cloned Cre recombinase into CPV2 and CPV14, creating CPV2/Cre and CPV14/Cre constructs, respectively. In these constructs, the mouse metallothionein gene is downstream of the Cre coding sequence, providing introns and a polyadenylation signal (Table 1) . In most cases, a tyrosinase minigene cassette was coinjected with the Cre transgenes to facilitate identification of transgenic animals by coat color. ⁴⁵  

Seven transgenic lines were established for each Cre transgene construct. Lines coinjected with tyrosinase minigene were also transgenic for this marker (Table 2) . The expression of Cre in these transgenic lines was tested by crossing the transgenic mice to a ROSA26 reporter mouse strain in which LacZ expression is dependent on Cre-mediated recombination. ⁴⁸ Once Cre-mediated recombination occurs in a given cell, that cell and all its descendants will express LacZ. Therefore, tissues that stained blue with x-gal–marked cell lineages in which Cre had been present and active. 

The results of these analyses are shown in Figure 3 and summarized in Table 2 . Of the seven CPV14/Cre transgenic lines, four lines (57%) showed evidence of LacZ expression in most lens epithelial cells in addition to the lens fibers, demonstrating enhanced lens epithelial activity of the CPV14 promoter (Figs. 3A 3B 3C 3D 3A′ 3B′ 3C′ 3D′) . Two lines (29%) demonstrated incomplete epithelial activity while maintaining extensive reporter expression in the fiber cells (Figs. 3E 3E′ 3F 3F′) . In the remaining transgenic line (14%), LacZ expression was limited in both the lens epithelium and fibers (Figs. 3G 3G′) . In our experience, the lack of blue staining in the deeper fiber cells represents a failure of staining reagents to diffuse deep into the intact lens, but not the absence of Cre expression in the embryonic fiber cells. These results contrasted sharply with those obtained from the CPV2/Cre transgenic lines. In four transgenic lines (57%), abundant LacZ reporter expression was detected in the lens fibers, but was largely absent from the lens epithelium (Figs. 3H 3I 3J 3K 3H′ 3I′ 3J′ 3K′ ; Table 2 ). One transgenic line (14%) showed sparse LacZ expression in the lens epithelial cells, despite extensive activity in the lens fibers (Figs. 3L 3L′) . In the remaining two transgenic lines (29%), LacZ expression was limited in both the lens epithelium and fiber cells (Figs. 3M 3M′ 3N 3N′) . Therefore, the insertion of this Pax6 consensus binding site expanded the expression domain of the αA-crystallin promoter to most of the lens epithelium. It is noteworthy that the majority of the Cre transgenic lines (11/12 lines, 92%) that were also transgenic for the coinjected tyrosinase minigene cassette, ocular expression of LacZ was not restricted to the lens. These ectopic sites of ocular transgene expression included retina, RPE, and cornea (Table 2) . Whereas in transgenic lines generated without tyrosinase minigene coinjection, reporter gene expression within the eye was limited to the lens (100%, two lines). 

The Extent of CPV14 versus CPV2 Promoter Activity in the Lens Epithelium

Though analyses of lens sections indicated the enhanced epithelial activity of CPV14 promoter, it is difficult to evaluate quantitatively the proportion of the entire lens epithelium stained with x-gal in tissue sections. To determine the efficiency of Cre activity in the lens epithelium, we analyzed LacZ reporter expression using wholemount lens epithelial sheets. For the CPV14/Cre construct, the MLR10 and MLR32 lines consistently demonstrated reporter expression in most lens epithelial cells (Figs. 4A 4B) . In contrast, the MLR39 and MLR37 transgenic lines made from the CPV2/Cre construct, showed little or no LacZ activity in the lens epithelium (Figs. 4C 4D) . This is consistent with the observed earlier onset of LacZ reporter expression (tested in MLR10) at E10.5 in CPV14/Cre transgenic mice versus between E11.5 and E12.5 (tested in MLR39) for CPV2 transgenic mice (described later). Overall, though transgene expression driven by the native murine αA-crystallin promoter was limited to the lens fiber cells, the addition of the Pax6 consensus binding sequence was capable of influencing the expression pattern of the −282/+43 murine αA-crystallin promoter to include the lens epithelium. 

Although the goal of this study was to determine the influence of regulatory elements on transgene expression within the eye, we were also aware that these experiments might generate transgenic lines that could be used to delete genes in lens cells conditionally at different stages of differentiation. Because the utility of these transgenic mice for lens-specific gene deletion depends on the developmental specificity of Cre expression, we evaluated the global pattern of transgene expression in developing embryos. For this purpose, we conducted wholemount x-gal staining of the embryos in several transgenic lines after crossing to the Cre reporter strain. For the CPV14/Cre construct, MLR10, MLR32, and MLR31 embryos were tested (Figs. 5A 5B 5C) . MLR39, MLR37, and MLR34 were chosen from transgenic lines made with the CPV2/Cre construct (Figs. 5D 5E 5F) . The staining patterns of these transgenic lines were characterized by positive staining in the embryonic lens. Although all transgenic embryos stained with x-gal in the eye, extraocular staining patterns were unique to each transgenic line and varied widely, probably due to different transgene integration sites. 

Lines MLR10 from the CPV14/Cre lines and MLR39 from the CPV2/Cre lines emerged as transgenic families in which there was good transgene expression in the lens and minimal extraocular transgene expression. Although no evidence of Cre expression was detected in the lens placode from either MLR10 or MLR39 embryos (data not shown), in MLR10 mice, Cre expression was readily and consistently seen at E10.5 during the lens pit–early lens vesicle stage (Fig. 6A) . Over the next 24 hours, evidence of Cre expression was detectable in the majority of both the lens epithelium and differentiating lens fibers (Fig. 6B) , and this expression expands even more by E12.5 (Fig. 6C) . Although there is some mouse-to-mouse variability, the pattern of expression in the lens is quite consistent. Occasionally, a few cells in the center of the cornea demonstrate evidence of Cre expression, but we have never seen Cre expression in the neural retina or RPE of MLR10 mice. Outside the eye, MLR10 expression can usually be detected in the snout and vibrissae follicles, as well as in parts of the midbrain and pituitary gland (Fig. 5A) . Very little extraocular expression was detected in MLR39 embryos. Within the MLR39 eye, Cre expression in the RPE was evident by E11.5 (Fig. 6D) , but lens expression of Cre was not evident at this stage. In the MLR39 lens, Cre activity was first evident at E12.5 when most of the fiber cells stain with x-gal (Fig. 6E) . Cre activity was only rarely detected in lens epithelial cells, but most lens cells initiated Cre expression subsequent to a fiber differentiation signal (Figs. 6E 6F) . Although the MLR39 RPE consistently stained with x-gal, blue-stained cells were not consistently visible in the neural retina (compare Figs. 6D and 6F with 6E ). 

Discussion

As expected, the −282/+43 fragment of the murine αA-crystallin promoter was effective in driving transgene expression only in lens fiber cells. Therefore, additional distal regulatory elements are likely to be required for the endogenous expression of the murine αA-crystallin gene in the lens epithelium. In contrast, the −214/+38 fragment of the murine αB-crystallin promoter consistently directed reporter gene expression to the lens epithelium and RPE and, in all but one case, failed to direct detectable reporter gene expression in the lens fiber cells. This pattern of expression was largely unaltered (with the exception of corneal epithelial expression in one founder) by the addition of the −282/−86 αA-crystallin promoter fragment and the αB-crystallin first intron and polyadenylation sequence. These elements were added to the αB-crystallin promoter in an attempt to increase the overall expression level of the transgene, but we reasoned that these modifications (particularly the inclusion of the αA-crystallin promoter fragment) might alter promoter specificity. The −282/−86 αA-crystallin promoter fragment contains a DE1 site −108/−100 that has been shown recently to bind c-Maf, a transcription factor essential for lens-specific expression of α-, β-, and γ-crystallins ⁵² , ¹⁷ but other regulatory elements of αA-crystallin promoter such as the αA-CRYBP1 site −66/−57, ^{18 20 53} the Pax6 site −49/−33 ¹⁷ , the TATA box -31/−26, ²¹ and additional proximal elements at -25/−12, ²¹ −7/+5, ⁵⁴ and +/24/+43 ^{22 53} are not included in the composite promoter. Therefore, the DE1 site was insufficient to activate the −214/+38 αB-crystallin promoter in the lens fiber cells, suggesting that the DE1 element must interact with other αA-crystallin promoter elements to confer expression in lens fiber cells. Consistent with this interpretation, the −287/−85 murine αA-crystallin promoter fragment did not work in concert with an SV40 promoter to activate transcription when transfected into lens epithelial explants. ⁵⁵  

In contrast, the insertion of a single copy of a Pax6 consensus binding site in either orientation at −86, between the DE1 site and the αA-CRYBP1 site of murine αA-crystallin promoter was able to alter the expression pattern of this promoter independently. During lens development, Pax6 expression exhibits distinct polarity, with higher level expression in the lens epithelium that gradually diminishes during fiber differentiation. ^{30 31} In our modified αA-crystallin promoter, there are two cis-acting elements that bind Pax6 and stimulate the promoter activity, one from the endogenous αA-crystallin promoter −49/−33, ¹⁷ and the other from the inserted Pax6 consensus binding site. Thus, the insertion of this perfect consensus Pax6 binding site enhanced transgene expression in the lens epithelium without altering the lens specificity of the endogenous αA-crystallin promoter sequences. At least two binding sites for Pax6 have been found in the lens epithelium-expressed chicken αA-crystallin and δ-crystallin genes as well as the murine αB-crystallin gene, each of which contribute to optimal promoter function. ^{4 34 35} Recent studies indicated that c-Maf could form a complex with the coactivator CBP/p300 to activate the αA-crystallin promoter, but Pax6 does not directly interact with this complex. ⁵⁶ Because c-Maf expression can be upregulated by Pax6, ⁵⁷ Pax6 can regulate αA-crystallin expression both directly and indirectly. The consensus Pax6 binding site inserted into the αA-crystallin promoter may recruit additional Pax6, which may act in concert with the c-Maf–CBP/p300 complex to activate transcription. Previous studies have suggested that Pax6 alone is not sufficient to activate the αA-crystallin promoter in the lens, despite the observation that the −49/−33 region of murine αA-crystallin gene is capable of binding Pax6 and stimulating promoter activity. ^{17 58 59} Mutations in both the DE1 site and αA-CRYBP1 site in the −116/+46 fragment abolished the function of the αA-crystallin promoter in transgenic mice, despite the presence of the native Pax6 binding sequence. ⁶⁰  

The −96/−76 region of the murine αA-crystallin promoter (where the consensus Pax6 site was inserted) has not been demonstrated to bind nuclear proteins and modulate promoter activity. In addition, replacement of the inserted Pax6 binding site with other cis regulatory elements in the modified murine αA-crystallin promoter did not lead to epithelial expression in transgenic mice, indicating that enhanced expression of transgene in the lens epithelium was specific to the Pax6 consensus binding site (Robinson ML, et al. IOVS 1997;38:ARVO Abstract 2694). Thus, we think it is unlikely that enhanced epithelial transgene expression resulted from disruption of a negative regulatory element by the Pax6 consensus insertion. The use of different reporters also demonstrated that the influence of the inserted Pax6 binding site was not dependent on particular sequences in the reporter genes. To date, we have not tested the consensus Pax6 site in any other part of the αA-crystallin promoter. 

The reason we used Cre as one of the reporter genes was twofold. In addition to being able to use a Cre reporter line to document Cre activity, we reasoned that the resultant mice might be useful for tissue-specific gene deletion in the lens. In several of the transgenic mice generated, ocular tissues other than the lens stained positively for x-gal in eyes of both CPV2/Cre and CPV14/Cre transgenic mice. Ocular Cre expression outside the lens was not seen in the transgenic lines where the tyrosinase minigene was absent. Therefore, we think that it is likely that the tyrosinase minigene was able to influence the activity of the coinjected CPV2 and CPV14 promoters within the eye. Transgene expression outside the eye varied considerably from line to line and was probably the result of integration site-specific effects. 

Transgenic lines MLR10 and MLR39 demonstrated different patterns of Cre expression within the lens and reasonably little extraocular Cre expression, and we believe that these mice represent potentially useful tools for conditional gene deletion experiments in the lens. Absolute tissue specificity is a rare characteristic of Cre-expressing transgenic mouse lines, and the usefulness of any particular Cre line must be evaluated in the context of the question being asked and the normal expression pattern of the gene one desires to delete. 

During the course of these studies, others independently generated the LeCre transgenic mouse line. ³² In the LeCre mice, the Pax6 P0 promoter, including the upstream ectoderm enhancer, was used to drive expression of Cre recombinase to the lens placode. Cre recombinase expression in the LeCre mice starts at the lens placode stage, making these mice particularly useful for deleting genes during early lens development. Cre expression at this early stage also explains the lack of lens specificity for the LeCre line, as the Cre-expressing tissue also gives rise to the corneal epithelium, conjunctiva and part of the eyelid. ³² Like the transgenic mice we describe in this report, the LeCre mice do not exhibit complete ocular specificity of Cre expression. LeCre mice consistently express the Cre transgene in the developing pancreas. Cre expression in our best CPV14/Cre transgenic line (MLR10) initiates during the lens pit stage and is detectable in most lens cells at the lens vesicle stage when lens fiber cell differentiation commences. Therefore, MLR10 may be useful in the study of lens fiber differentiation, provided the limited extraocular Cre expression does not interfere with embryonic survival or with the development of tissues that indirectly influence the eye. We anticipate that conditional targeting of genes with different Cre transgenic lines will provide valuable functional information on specific genes at different stages of lens development. 

 Supported by National Eye Institute Grants EY06511 (MLR), EY12995 (MLR), EY10448 (PAO), and EY10803 (PAO); and the Columbus Children’s Research Institute and NCI The Comprehensive Cancer Center at The Ohio State University.

 Submitted for publication August 7, 2003; revised December 26, 2003; accepted January 29, 2004.

 Disclosure: H. Zhao, None; Y. Yang, None; C.M. Rizo, None; P.A. Overbeek, None; M.L. Robinson, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Structure of Promoter Constructs Used to Generate Different Transgenic Mouse Lines

Name of Construct	Promoter	Transgene	Intronic Sequence	Polyadenylation Sequence
CPV2/hGH	−282/+43 of mouse αA-crystallin gene	Human growth hormone (hGH)	hGH	hGH
αB1/hGH	−214/+38 of mouse αB-crystallin gene	hGH	hGH	hGH
αA/αB4/hGH	−282/−86 mouse of αA-crystallin fused to −214/+38 of mouse αB-crystallin gene	hGH	The first intron of mouse αB-crystallin gene and hGH	3′ Untranslated and polyadenylation sequences of mouse αB-crystallin gene
CPV14/hGH	−282/+43 of mouse αA-crystallin gene with an inserted copy of Pax6 consensus binding site in a 5′-3′ orientation at −86	hGH	hGH	hGH
CPV15/hGH	Same as CPV14/hGH except the Pax6 consensus binding site is inserted in an opposite orientation	hGH	hGH	hGH
CPV14/Cre	−282/+43 of mouse αA-crystallin gene with an inserted copy of Pax6 consensus binding site in a 5′-3′ orientation at −86	Cre recombinase	Mouse metallothionein gene	Mouse metallothionein gene
CPV2/Cre	−282/+43 of mouse αA-crystallin gene	Cre recombinase	Mouse metallothionein gene	Mouse metallothionein gene

Figure 1.

 

Expression analysis of human growth hormone transgenes driven by α-crystallin promoters. (A) Transgenic constructs with the α-crystallin promoter fragments used to drive the expression of an hGH reporter gene. The CPV2/hGH construct contained the −282/+43 murine αA-crystallin promoter (heavily stippled box) driving hGH (filled box). The αB1/hGH differed from CPV2/hGH only by the replacement of the murine αA-crystallin promoter with the −214/+38 αB-crystallin promoter (open box). The αA/αB4/hGH construct contained a chimeric promoter consisting of the −282/−86 fragment of the murine αA-crystallin promoter (heavily stippled box) fused to the −214/+38 murine αB-crystallin promoter (open box). The chimeric promoter was followed by the first intron of the murine αB-crystallin gene (striped box). The 3′ untranslated region of the murine αB-crystallin cDNA (checkered box) followed the stop codon for hGH in this construct. Transgene expression was analyzed both by in situ hybridization (B, C, F, G, J, K) and immunohistochemistry (D, E, H, I, L, M) on E15.5 transgenic founder embryos injected with CPV2/hGH (B–E), αB1/hGH (F–I), or αA/αB4/hGH (J–M). Boxed regions in (B, F, J, D, H, L) are shown at higher magnification in (C, G, K, E, I, M). Bright-appearing silver grains reveal transgene expression in dark-field micrographs. Immunohistochemistry localized hGH expression (brown) in hematoxylin-counterstained eye sections. Sites of transgene expression, depending on the construct, included the lens epithelium (e) and fiber cells (f) as well as the retinal pigment epithelium (rpe) and cornea (c). Scale bars, 50 μm.

Figure 2.

 

The addition of a Pax6-binding site into the murine αA-crystallin promoter enhanced transgene expression in the lens epithelium in an orientation-independent manner. (A) Modified αA-crystallin promoter transgenic constructs. A single copy of Pax6 consensus binding site (striped box) was inserted at position −86 of the murine αA-crystallin promoter (heavily stippled box) to create the CPV14 promoter. The CPV15 promoter differed from CPV14 only by the orientation of the Pax6 consensus binding site. The sequence of the inserted fragment was surrounded by the endogenous αA-crystallin promoter sequences (smaller bold font) and the Pax6 binding sequence is shown within the filled box. hGH (filled box) was placed downstream of the modified αA-crystallin promoters. In situ hybridization (B, C, F, G) and immunohistochemistry (D, E, H, I) were performed on E15.5 embryos transgenic for CPV14/hGH (B–E) or CPV15/hGH (F–I) to reveal transgene expression. Boxed regions in (B, F, D, H) are shown at higher magnification in (C, G, E, I). For in situ hybridization, transgene-specific signals were revealed by reflected light in dark-field pictures. For immunohistochemistry, brown staining signals indicated the expression of transgene product. e, lens epithelium; f, lens fiber cells. Scale bars, 50 μm.

Table 2.

 

Cre Expression in Ocular Tissues of Different Transgenic Mouse Lines

Constructs	Transgenic Line	Expression in Lens Epithelium	Expression in Lens Fiber Cells	Ectopic Expression within the Eye	Tyrosinase Coinjection
CPV14/Cre	MLR10	+++	+++	No	No
	MLR32	+++	+++	Retina, RPE cornea	Yes
	MLR31	+++	+++	Retina, ciliary body	Yes
	MLR30	+++	+++	Retina, ciliary body	Yes
	MLR33	++	+++	RPE	Yes
	MLR28	++	+++	Retina	Yes
	MLR29	+	+	No	Yes
CPV2/Cre	MLR53	−	+++	No	No
	MLR39	−	+++	RPE, ciliary body	Yes
	MLR38	−	+++	Retina, RPE, ciliary body, cornea	Yes
	MLR36	−	++	Retina, RPE	Yes
	MLR37	+	+++	Retina, RPE, cornea	Yes
	MLR35	+	+	Retina, RPE	Yes
	MLR34	+	+	Retina, RPE	Yes

 The intensity of x-gal staining was visually estimated as being weak (+), medium (++), strong (+++) or not detectable (−).

Figure 3.

 

Analyses of LacZ reporter expression in eyes from Cre transgenic lines. After crossing to the ROSA26 reporter strain, neonatal eyes heterozygous for both the Cre transgenes and the ROSA26 reporter allele were analyzed for LacZ expression (blue), reflecting the activity of Cre recombinase in transgenic mice. The CPV14/Cre transgenic lines were MLR10 (A, A′), MLR32 (B, B′), MLR31 (C, C′), MLR30 (D, D′), MLR33 (E, E′), MLR28 (F, F′), and MLR29 (G, G′). The CPV2/Cre transgenic lines were MLR53 (H, H′), MLR39 (I, I′), MLR38 (J, J′), MLR36 (K, K′), MLR37 (L, L′), MLR34 (M, M′), and MLR35 (N, N′). The staining pattern is shown in low magnification (A–N) to show total ocular reporter expression, and the boxed regions are shown in high magnification (A′–N′) to distinguish lens epithelial (arrowhead) from lens fiber ( ) staining. Sections were counterstained with nuclear fast red. Scale bars, 200 μm.

Figure 4.

 

Detection of LacZ expression in lens epithelial wholemounts from mice transgenic both for Cre and the ROSA26 reporter gene. Lens capsule and epithelium was removed from transgenic mice at approximately P21 and stained for β-galactosidase activity. The samples were counterstained with hematoxylin to visualize the nuclei. Lens epithelial wholemounts from CPV14/Cre transgenic mouse lines MLR10 (A) and MLR32 (B), as well as CPV2/Cre transgenic lines MLR39 (C) and MLR37 (D), were examined. Nontransgenic lenses (E) and lenses transgenic only for the ROSA26 reporter gene (F) were used as the negative control. Black arrow: sporadic positive staining of the MLR37 transgenic line; white arrows: sporadic negative staining of the MLR10 and MLR32 transgenic lines. Scale bar, 100 μm.

Figure 5.

 

Extraocular Cre expression in CPV2/Cre and CPV14/Cre transgenic lines. Global Cre reporter expression was revealed by x-gal staining (blue) of embryos heterozygous for the Cre transgene and ROSA26 reporter allele. Wholemount embryos from CPV14/Cre transgenic mouse lines MLR10 (A), MLR32 (B), and MLR31 (C) and the CPV2/Cre transgenic lines MLR39 (D), MLR37 (E), and MLR34 (F) are shown. All embryos were from E12.5 except MLR34, in which E14.5 embryos were stained. Arrows: strong positive x-gal staining in the embryonic lenses from different transgenic lines. Extraocular reporter expression varied in independently generated transgenic lines independent of whether CPV2/Cre or CPV14/Cre constructs were used.

Figure 6.

 

Expression of Cre recombinase during lens development in transgenic lines made with CPV14/Cre and CPV2/Cre constructs. Transgenic lines MLR10 (A–C) and MLR39 (D–F) were chosen for analysis of Cre transgene expression during lens development by x-gal staining after crossing with the Rosa26 reporter mouse strain. Developmental stages studied included E10.5 (A), E11.5 (B, D), E12.5 (C, E), and E14.5 (F). In the MLR10 line, Cre transgene expression was observed as early as E10.5 and expanded to most of the lens cells by E12.5. In the MLR39 line, Cre expression was not detected until E12.5, when it was largely restricted to the lens fiber cells. Arrows: transgene expression in the RPE (D–F). Scale bars, 100 μm.

Supplementary Materials

Appendix - (PDF) 

The authors thank Kumud Majumder, Lixing Reneker, and John Ash for many helpful discussions and encouragement during the initiation of this project; Long Vien, Barbara Harris, Janet Weslow-Schmidt, and Brenda Van Dyke for technical assistance; and Heithem El-Hodiri, Carlton Bates, and Nicolas Berbari for critical review of the manuscript.