Inhibition of Crystallin Expression and Induction of Apoptosis by Lens-Speci®C E1A Expression in Transgenic Mice

Inhibition of crystallin expression and induction of apoptosis by lens-speci®c E1A expression in transgenic mice

Qin Chen1, John D Ash1,4, Phil Branton2, Larry Fromm3 and Paul A Overbeek*,1

1Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, Texas, TX 77030, USA; 2Department of Biochemistry, McGill University, Montreal, Quebec, Canada, H3G1Y6; 3Skirball Institute, New York University Medical School, New York, NY 10012, USA

Previous studies have shown that the adenovirus E1A (Piatigorsky, 1981). The cells located at the posterior oncoprotein can bind to and inactivate the retinoblasto- surface of the lens vesicle are induced to dierentiate ma tumor suppressor protein (pRb) and the transcrip- into primary lens ®ber cells that elongate to ®ll the tional coactivators CBP/p300. In this study, wild-type lens vesicle. Growth of the lens throughout life E1A12S or two deletion mutants (delN, which binds pRb involves proliferation within the epithelial cell popula- but not CBP/p300; delCR2, which binds to CBP/p300 tion followed by the induction of dierentiation into but not pRb) were linked to the lens-speci®c aA- secondary ®ber cells speci®cally around the lateral crystallin promoter, and used to generate transgenic margins of the lens. Once ®ber cell dierentiation is mice. Lens ®ber cells expressing E1A12S or delCR2, initiated, a number of cellular responses occur both of which bind to CBP/p300, failed to upregulate b- including upregulation of the cell cycle inhibitor crystallin and g-crystallin expression. In contrast, lens Kip2 (p57) (Zhang et al., 1998) and activation of ®ber cells expressing delN showed signi®cant expression expression of ®ber cell-speci®c proteins which include of b- and g-crystallins. Lens ®ber cells expressing delN the b- and g-crystallins (McAvoy, 1978; Treton et al., showed cell cycle entry, marked apoptosis, and evidence 1991). These changes in gene expression are presumed for p53 activation, while cells expressing either 12S or to be the result of alterations in the activities of delCR2 showed limited apoptosis and no evidence for speci®c transcription factors. Targeted mutations in upregulation of the p53-inducible gene p21. Our results the genes encoding c-Maf, Sox1, and Prox1 have all suggest that the transcriptional coactivators CBP and/or resulted in speci®c defects in ®ber cell dierentiation p300 are required for the dramatic increases in crystallin (Ring et al., 2000; Nishiguchi et al., 1998; Wigle et al., expression that accompany terminal dierentiation in the 1999). In the present study we wanted to assess lens, and also for activation of p53 in response to possible roles for the transcriptional co-activators inactivation of pRb in the lens. CBP/p300 in ®ber cell dierentiation. Oncogene (2002) 21, 1028 ± 1037. DOI: 10.1038/sj/onc/ CBP (CREB-binding protein) and p300 are distinct 1205050 but functionally related co-activator proteins. Both proteins are E1A-binding proteins and they share Keywords: E1A; CBP/p300; crystallin; apoptosis; trans- common protein domains (Eckner et al., 1994). It has genic mice been shown that CBP and p300 can interact with numerous sequence-speci®c transcription factors, including CREB (Chrivia et al., 1993), c-Fos (Bannister and Introduction Kouzarides, 1995), E2F-1 (Trouche and Kouzarides, 1996) and p53 (Gu and Roeder, 1997). Both CBP and The vertebrate lens provides an attractive model p300 are required for normal embryogenesis. Mouse system in which to study the links between cell cycle embryos lacking p300 or CBP show early embryonic exit and dierentiation-speci®c gene expression. Dur- lethality (Yao et al., 1998). In addition, microinjection of ing embryonic development, the ocular lens is initially CBP and p300 antibodies into myoblasts can block formed as an invagination of the surface ectoderm. As terminal dierentiation (Eckner et al., 1996). These dierentiation proceeds, the lens forms a hollow results indicate that CBP and p300 are essential for sphere of epithelial cells termed the lens vesicle multiple aspects of gene regulation, but may be particularly important for cellular dierentiation. The adenoviral E1A protein can bind to and *Correspondence: PA Overbeek, Department of Molecular and inactivate both pRb and CBP/p300 (Moran, 1993; Cellular Biology, Baylor College of Medicine, One Baylor Plaza, Jones, 1995). Via these interactions, E1A simulta- Houston, TX 77030, USA; E-mail: [email protected] neously relieves growth suppression and inhibits 4Current address: Department of Ophthalmology, Dean A. McGee cellular dierentiation. E1A encodes two major Eye Institute, 608 Stanton L. Young Blvd., Oklahoma City, OK mRNA species, 12S and 13S, which are generated 73104, USA Received 2 July 2001; revised 19 September 2001; accepted 9 through alternative splicing of a single transcript. October 2001 There are three regions in the E1A12S protein which E1A inhibits crystallin expression by binding CBP/p300 Q Chen et al 1029 are required for transformation: (1) the N-terminus Expression of CBP and p300 in normal lens (amino acids: 2 ± 24); (2) conserved region 1 (CR1, amino acids: 40 ± 80); and (3) conserved region 2 RT ± PCR was performed with lens, brain and kidney (CR2, amino acids: 120 ± 140) (Barbeau et al., 1994). RNA from newborn non-transgenic (FVB) mice to The CBP/p300 proteins bind the N-terminus and assay for expression of CBP and p300. We found that CR1 of E1A, while pRb family members bind CR1 CBP and p300 are both expressed in the lens at levels and CR2 (Barbeau et al., 1994). E1A binding to comparable to brain and kidney (Figure 1b). There- pRb causes the release of free E2F transcription fore, either protein (or both) may be the relevant target factors, leading to initiation of DNA synthesis for E1A in the lens. (Bagchi et al., 1990). In addition, E1A mutants that interact with CBP/p300, but not pRb proteins, are Lens histology capable of severely perturbing dierentiation (Mym- ryk et al., 1992), and repressing tissue-speci®c gene The transgenic lenses were smaller and often hollow expression. due to degeneration of the primary ®ber cells in the In this study we have generated transgenic mice that center of the lens (Figure 1d ± f). Secondary ®ber cells express full-length E1A12S and two mutant forms of showed changes in cell morphology, including defects E1A in the lens. One mutant, which we term delN (for in cellular elongation. There was a signi®cant increase deletion of N-terminus), has lost a segment of the in the number of cells at the posterior of the lens, CBP/p300 binding domain (deletion 1101 in Barbeau et suggesting inappropriate cellular proliferation, espe- al., 1994). The other, which we term delCR2 (for cially in 12S (Figure 1d,h) and delN lenses (Figure deletion of conserved region 2), has lost a portion of 1e,i). The lens histology in delCR2 was mosaic (Figure the pRb binding domain (deletion 1108 in Barbeau et 1f,j). al., 1994). By comparing the eects of these three BrdU incorporation assays have shown previously variants of E1A we assess whether CBP and/or p300 (Fromm et al., 1994) that proliferation in a non- are required for diverse aspects of ®ber cell dierentia- transgenic lens is con®ned to the epithelial cells (Figure tion. Our results suggest that CBP/p300 activity is 2e). In contrast, in 12S transgenic embryos an average required for induction of expression of the ®ber-speci®c of 33% of the cells at the posterior of the lens were crystallins (b- and g-) during normal dierentiation of BrdU positive (Figure 2f) (Table 2), while in delN the lens. In addition, CBP/p300 function is apparently transgenic embryos an even higher percentage of cells required for p53 activation in ®ber cells that are (60%) at the posterior of the lens were BrdU positive stimulated to abnormally enter the cell cycle by (Figure 2g) (Table 2). In comparison, delCR2 trans- inactivation of pRb. genic lenses showed a low but signi®cant percentage (5%) of BrdU positive ®ber cells (Figure 2h) (Table 2). In order to assay for changes in cell cycle genes, in situ hybridizations were done using probes for expression Results of E2F1 and cyclins A2, B1, D3 and E. We found that these genes were all upregulated in both 12S (Figure Expression of E1A in lenses of transgenic mice 3b,e,h,k,n) (Table 1) and delN (Figure 3c,f,i,l,o) (Table The mouse aA-crystallin promoter was linked to the 1) ®ber cells at the posterior of the lenses. These results E1A coding sequences (Figure 1a). Newborn mice are consistent with the prediction that both 12S and transgenic for the 12S and delCR2 constructs delN can bind pRb and inactivate pRb function. One typically died before weaning with visible skin consequence of pRb inactivation would be release of problems. We believe, but have not experimentally E2F transcription factors, which are then able to con®rmed, that these skin defects are caused by upregulate E2F1 and cyclin family members (A2, B1, ectopic expression of the transgenes (see Figure 2b D3 and E), leading to inappropriate cell proliferation, and data not shown), and that the skin defects, as previously described (Fromm et al., 1996). The rather than the lens defects, are the cause of the expression of these genes in the 12S ®ber cells suggests premature mortality. Six 12S and two delCR2 that E2F transcription factors can function in the founder transgenic embryos, as well as two delN absence of CBP/p300 activity. embryos from two transgenic families were character- ized in detail (Tables 1 and 2). Most of the embryos Expression of fiber cell differentiation markers showed alterations in lens histology that aected the ®ber cells at the posterior of the lens (Figure 1). The Lens development is accompanied by a well-de®ned histology of the transgenic lenses is discussed in more temporal and spatial pattern of expression of a, b and g- detail in a later section. In situ hybridization showed crystallins (McAvoy, 1978). The a-crystallins are that transcripts of both E1A12S and the deletion produced by epithelial and ®ber cells, while b- and g- mutants were present speci®cally in cells at the crystallins are ®ber cell speci®c. In situ hybridizations posterior of the lens with little or no expression in (Table 1, Figure 4) and immunohistochemistry (IHC) lens epithelial cells (Figure 2). One founder delCR2 (Table 2, Figure 5) were used to assay for the expression embryo showed a mosaic pattern of transgene of b- and g-crystallins. In FVB lenses, b- and g-crystallin expression in the lens (Figure 2d). transcripts (Figure 4a,h) (Table 1) and proteins (Figure

Oncogene E1A inhibits crystallin expression by binding CBP/p300 Q Chen et al 1030

Figure 1 (a) Diagrams of the microinjected constructs, indicating the aA-crystallin promoter, E1A coding sequences, and the SV40 intron/polyA region. (b) RT ± PCR analysis of CBP and p300 expression in brain, kidney and lens. Lane 1: marker; Lanes 2, 3, 4: p300; Lanes 6, 7, 8: CBP; Lanes 5, 9: RT omitted. (c ± j) Histology of representative lenses at E15.5. In contrast with non-transgenic FVB lenses (c ± g), all of the transgenic lenses (d ± f, h ± j) show disruption of ®ber cell elongation, degeneration of primary ®ber cells in the core of the lens, and an abundance of nuclei at the back of the lens, especially in 12S (d,h) and delN (e,i). The lens phenotype in delCR2 (f,j) was mosaic. (g ± j) Higher magni®cations of the regions within the rectangles in c ± f. Abbreviations: co, cornea; le, lens epithelium; lf, lens ®bers; nr, neuronal retina. Scale bars=500 mm

5a,e) (Table 2) were detected speci®cally in lens ®ber inhibition of expression of b-andg-crystallins by E1A, cells, as expected. Interestingly, in 12S (Figure 4b,e,i,l; and imply that this inhibition requires the region that Figure 5b,f) transgenic lenses, transgene-expressing ®ber binds to CBP/p300, and is not due simply to loss of cell cells at the posterior of the lenses (Tables 1 and 2) cycle control. showed little or no expression of b- and g-crystallins. In the mosaic delCR2 (Figure 4d,g,k,n; Figure 5d,h) lenses, c-Maf expression ®ber cells with strong expression of b-andg-crystallin are intermingled with cells showing little or no expression at Gene targeting studies have shown that c-Maf is the posterior of the lenses. By comparison, in delN required for crystallin expression (Ring et al., 2000). (Figure 4c,f,j,m; Figure 5c,g) transgenic embryos, b-and To ascertain whether inhibition of CBP/p300 by E1A g-crystallin expression both showed a modest reduction, results in a loss of c-Maf expression, in situ but there was clear evidence for continued expression in hybridizations were performed. In nontransgenic ®ber cells at the posterior of the lenses (Tables 1 and 2). (Figure 6a) and transgenic lenses (Figure 6b ± d), c- These results show that there is an almost complete Maf transcription was upregulated at the onset of ®ber

Oncogene E1A inhibits crystallin expression by binding CBP/p300 Q Chen et al 1031

Figure 2 Patterns of E1A transgene expression and BrdU incorporation. Expression of E1A (white dots) was assayed by in situ hybridization in E15.5 embryos. There was no expression of E1A in non-transgenic FVB lenses (a). E1A expression was localized speci®cally to the posterior lens cells in transgenic lenses (b,c,d). In 12S (b), expression of the transgene in the skin was detected (arrow). The expression pattern of delCR2 (d) was mosaic. BrdU incorporation was assayed by immunohistochemistry. In the FVB lens (e), BrdU incorporation (brown nuclear stain) is restricted to the epithelial (e) cells. 12S (f) and delN (g) transgenic lenses showed multiple BrdU positive cells at the back of the lens. In delCR2 lenses (h) only a few BrdU positive ®ber cells were observed. Scale bars=500 mm

Table 1 Gene expression in ®ber cells (E15.5) (in situ hybridization results) Transgenic Number of integration Probes constructs sites assayed E1A E2F1 Cyclin E Cyclin A2 Cyclin B1 Cyclin D3 b-Cryst. g-Cryst. p57 C-maf p53 Bax p21

E1A 12S 6 ++ ++ ++ ++ ++ ++ 77+++++7 E1AdelN 2 +++ ++ +++ +++ +++ +++ ++ + ++ ++ ++ ++ ++ E1AdelCR2 2 7/++ nt nt nt nt nt 7/++ 7/++ ++ ++ nt nt 7 FVB 77 7 7 7 + +++ +++ ++* ++* 777

Expression levels in cells at the posterior of the lens: 7: background level of expression; 7/++: mosaic expression; ++: modest or localized expression; +++: high expression and widespread expression. nt: not tested; *: expression in equatorial ®ber cells

Table 2 Fiber cell features (E15.5) Transgenic Number of integration b-Crystallina g-Crystallina TUNELb BrdUb constructs sites assayed (IHC) (IHC) (%) (%)

E1A 12S 6 +/7 +/7 935 (+1.6) (+3.1) E1AdelN 2 ++ ++ 20 60 (+4.1) (+6.3) E1AdelCR2 2 7/++ 7/++ 0.5 5 (+0.2) (+0.1) FVB +++ +++ 0 0 aImmunohistochemistry (IHC) scale (brown staining in cells at the posterior of the lens): +/7: lens ®ber cells very weakly stained; 7/++: mosaic combination of 7 and ++ expression levels; ++: ®ber cells all stained, but some less strongly than wild-type; +++: ®ber cells strongly stained (wild-type level). bPercentage of positive cells: the number of brown nuclei in posterior lens ®ber cells compared to the total number of nuclei in the same region as determined by Methyl Green or hematoxylin staining (standard deviation in parentheses)

cell dierentiation in the equatorial zone. Expression of Cell cycle regulation c-Maf was clearly maintained in the transgenic ®ber cells (Figure 6b ± d), indicating that neither inhibition Expression of p57 is required for the cell cycle exit of CBP/p300 nor loss of cell cycle control results in that accompanies ®ber cell dierentiation (Zhang et loss of c-Maf expression, and that the loss of b- al., 1998). Since 12S and delN expression are sucient crystallin expression is not due to inhibition of c-Maf to induce BrdU incorporation in ®ber cells (Figure 2), transcription. we assayed the transgenic lenses for loss of p57

Oncogene E1A inhibits crystallin expression by binding CBP/p300 Q Chen et al 1032

Figure 3 Changes in expression of E2F1, cyclins A2, B1, D3 and E. In situ hybridizations were performed. At E15.5, the expression of E2F1 and cyclin family members (A2, B1, D3, E) is upregulated in cells expressing 12S (b,e,h,k,n) or delN (c,f,i,l,o). Only the posterior regions of the lenses are shown. Scale bars=500 mm

expression by in situ hybridization. In non-transgenic that p53 can induce cell death by upregulating Bax, lenses, p57 expression is highly upregulated in even in the absence of CBP/p300 activity. In the non- secondary ®ber cells at the equatorial zone and transgenic lenses (Figure 7e), no lens-speci®c tran- declines in mature ®bers (Figure 6e). In the transgenic scripts of p21 were observed. However, there was a lenses, p57 was still induced at the equatorial zone, signi®cant increase in p21 expression in lens ®ber cells and was not repressed in transgene expressing cells expressing delN (Figure 7g). In contrast, 12S (Figure (Figure 6f ± h; Table 1). These results suggest that p57 7f) and delCR2 (Figure 7h) lenses showed no evidence expression is not regulated directly by either pRb or of p21 upregulation, indicating that CBP/p300 might CBP/p300, and is also not inhibited by re-entry into be required for p53-mediated induction of p21. the cell cycle.

Discussion Apoptosis in the lens To test whether E1A-mediated inactivation of pRb We have used the ocular lens in transgenic mice as a induces programmed cell death in lens ®ber cells, model system to provide evidence that CBP and/or TUNEL assays were performed. Apoptotic nuclei were p300 are essential factors for high level expression of prevalent in posterior lens cells expressing delN (Figure certain dierentiation-speci®c genes. Our results show 7c) as expected. In contrast, there were signi®cantly that the CBP/p300 interaction domain of E1A can fewer posterior apoptotic nuclei in secondary lens ®ber inhibit the expression of b- and g-crystallins (Tables 1 cells expressing 12S (Figure 7b), and apoptotic nuclei and 2). Thus CBP/p300 activity is apparently required were almost undetectable in lens ®ber cells expressing for normal lens ®ber cell dierentiation, and also for delCR2 (Figure 7d). The average percentage of certain aspects of p53 activation after inappropriate apoptotic secondary ®ber cells was 20% for delN cell cycle entry. A model depicting putative roles for versus 9% for 12S and 0.5% for delCR2 (Table 2). CBP/p300 in the lens dierentiation program is These results suggest that CBP/p300 activity is involved provided in Figure 8. in the apoptotic response in the secondary ®ber cells, In the ocular lens, when pRb was speci®cally perhaps being required for activation of p53. In situ inactivated by expression of the N-terminus of SV40 hybridizations were used to assay for expression of p53 T antigen, b-andg-crystallins were still expressed even and p53-regulated genes Bax and p21. We found that in the presence of inappropriate cell proliferation transcripts of p53 and Bax were increased in both 12S (Fromm and Overbeek, 1996). The transgenic lenses (Figure 7j,m) and delN (Figure 7k,n) compared to non- expressing delN show a comparable phenotype. In transgenic lenses where transcripts were not detected comparison, the proliferating ®ber cells in 12S (Figure 7i,l; Table 1). These results suggest that transgenic lenses show almost no expression of b- inactivation of pRb by E1A can upregulate p53, and and g-crystallins (Figures 4 and 5). The delCR2

Oncogene E1A inhibits crystallin expression by binding CBP/p300 Q Chen et al 1033

Figure 4 Loss of b- and g-crystallin transcription. Transcripts of b- and g-crystallin were assayed by in situ hybridizations. At E15.5, in non-transgenic lenses (a,h), b- and g-crystallin are highly expressed in lens ®ber cells. In 12S lenses (b,e,i,l), the E1A- expressing cells (brackets in e,l) at the posterior of the lens show little or no b-crystallin (b,e)org-crystallin (i,l) transcription. Crystallin transcripts were present in the degenerating primary ®ber cells in the center of the lens. In the mosaic delCR2 lenses (d,g,k,n), ®ber cells with strong expression (short arrows) are intermingled with cells showing little or no expression (long arrows). The proliferating cells at the back of the delN lenses (c,f,j,m) were still transcribing the b-crystallin, and more weakly the g-crystallin genes. e ± g and l ± n are higher magni®cations of the boxed regions in b ± d and i ± k respectively. Abbreviations: le, lens epithelium; nr, neuronal retina. Scale bars=500 mm transgenic lenses do not show widespread ®ber cell cells, it is possible that the high levels of crystallin proliferation, but still appear to lack expression of expression require chromatin decondensation, and the these dierentiation markers in transgene-expressing HAT activity of CBP/p300 may be required to cells, implying that E1A12S and delCR2 repress promote an open chromatin state. crystallin expression as a result of binding to CBP/ Various downstream target genes of p53 have been p300. E1A also actively represses muscle dierentia- identi®ed, including Bax, p21, MDM2, GADD45, and tion, at least in part, by targeting pRb and CBP/p300 cyclin G (Lill et al., 1997; Smith et al., 1997). E1A has (Eckner et al., 1996). previously been shown to repress p53-mediated activa- The transcription factors which drive expression of tion of the p21 and Bax promoters and to reverse p53- b-andg-crystallin genes are not unambiguously mediated cell cycle arrest and apoptosis in vitro (Lill et de®ned. However, likely candidates include c-Maf al., 1997). Mutation of the CBP/p300-interacting domain (Ring et al., 2000) and Sox1 (Nishiguchi et al., 1998). in E1A results in loss of inhibition of p53 (Somasundar- It has been demonstrated that c-Maf directly binds to am and El-Deiry, 1997). Our in vivo transgenic results are regulatory sequences in many if not all of the b- not entirely consistent with these previous in vitro crystallin genes (Ring et al., 2000). Our data show that studies. The consequence of inactivation of pRb in lens c-Maf expression in the lens is not sucient to activate ®ber cells is that the cells undergo p53-dependent cell b-crystallin expression. Instead, the activation of the b- death (Fromm et al., 1994; Morgenbesser et al., 1994). crystallin genes by c-Maf may require CBP/p300 co- Induction of apoptosis (in lens ®ber cells) was also transactivation. CBP and p300 possess intrinsic histone observed in E1A12S and delN transgenic lenses. Our in acetyltransferase (HAT) activity (Ogryzko et al., 1996). situ results show that p53 transcription was upregulated Since crystallins are highly expressed genes in lens ®ber when pRb was inactivated by E1A. Bax transcription

Oncogene E1A inhibits crystallin expression by binding CBP/p300 Q Chen et al 1034

Figure 5 Decreased levels of b- and g-crystallin proteins. Immunohistochemistry was used to detect b-(a ± d) and g-crystallin (e ± h) proteins. In non-transgenic lenses (a,e), both proteins are expressed (brown staining) in lens ®ber cells. In 12S lenses (b,f), the cells at the posterior of the lens (which are the cells expressing the transgene) do not express these dierentiation markers. In the mosaic delCR2 lenses (d,h), ®ber cells with protein expression (white arrow head) are intermingled with cells showing little or no detectable protein (black arrow head). The lens cells expressing delN (c,g) exhibited a modest reduction in the levels of these ®ber cell speci®c proteins. Scale bars=500 mm

Figure 6 Expression of c-maf and p57. In situ hybridizations for c-maf expression (red dots) are shown in a ± d. At E15.5, high levels of c-maf expression are observed in the equatorial zone (ez) in non-transgenic FVB lenses (a). Maf expression is still detectable in the cells at the posterior of the transgenic lenses (b ± d). Analogous in situ hybridizations for p57 expression (red dots) are shown in e ± h. Newly dierentiating post-mitotic cells in the equatorial zone (ez) of non-transgenic lenses (e) express high levels of p57. The transgenic lenses showed p57 expression throughout the cells at the posterior of the lens (f ± h). Scale bars=500 mm

was also upregulated. Although we have not demon- Many genes (like the b-andg-crystallins) are strated that the Bax expression is p53-mediated, our expressed at high levels only in terminally dieren- results suggest that p53 may induce Bax expression even tiated cells. Such genes may have a common in the absence of CBP/p300 activity. In contrast, the mechanism for gene activation, and may use tissue- expression of p21, a cyclin-dependent kinase inhibitor speci®c transcription factors (such as maf) to target whose expression is upregulated by p53 (el-Deiry et al., CBP/p300 co-activators to their promoters. Since both 1993), was barely detectable in 12S or delCR2 transgenic CBP and p300 have histone acetyltransferase activity lenses, even though p21 was obviously induced in the (Ogryzko et al., 1996), histone acetylation may be a delN transgenic lenses. These results suggest that CBP/ common mechanism for accomplishing cell type- p300 activity is required for p53 activation of p21 speci®c chromatin remodeling and gene activation. expression. Further experiments will be needed to clarify Our studies with E1A expression in transgenic lenses whether speci®c aspects of p53 activity in vivo are CBP/ provide in vivo data in a mammalian system to p300 independent. support this model.

Oncogene E1A inhibits crystallin expression by binding CBP/p300 Q Chen et al 1035

Figure 7 Apoptosis in the lens and expression of p53, Bax and p21. TUNEL assays (a ± d) were used to examine cell death in lenses from E15.5 embryos. Brown staining indicates DNA fragmentation, a characteristic feature of apoptosis. In the posterior cells of the delN lens there was extensive apoptosis (c). The 12S lenses (b) also showed TUNEL positive cells, while very few apoptotic cells were detectable in delCR2 (d) and FVB (a) lenses. Expression of p21 was analysed by in situ hybridization (e ± h). In the lens, p21 expression was detected only in the delN embryos (g). Expression of p53 and Bax were assayed by in situ hybridizations. In non-transgenic lenses (i,l), there was no detectable expression of p53 or Bax in ®ber cells. However, transcription of both genes was detected in 12S (j,m) and delN (k,n) lenses. Scale bars=500 mm

Materials and methods RT ± PCR RNA was isolated from lenses, kidney and brain of newborn Generation of the constructs and transgenic mice FVB mice using RNA Stat-60 (Tel-test B, Inc.). RT reactions HindIII ± PstI fragments encoding E1A12S, delN, or delCR2 were subsequently ampli®ed by PCR with sense primers (5'- were cloned into the crystallin promoter vector CPV2 GAGTCTGCTAATAGCAGG for CBP; 5'-AGAAGCG- (Robinson et al., 1995). Mutant delN has had the coding CAACGTCATCC for p300) and antisense primers (5'- sequences for residues 4 ± 25 deleted and therefore binds pRb TTCCTGACCCTGTGAAAC for CBP; 5'-CATGTTGTT- but not CBP/p300, while mutant delCR2 has had the CR2 CACAGACCG for p300) to generate a 550 bp fragment region deleted so that it no longer binds pRb (Barbeau et al., from CBP mRNA and a 540 bp fragment from p300 mRNA. 1994). The resulting plasmids (Figure 1a) were digested with EagI to release 2.2 kb fragments, which were isolated by In situ hybridization electrophoresis through a 1% agarose gel, and puri®ed by Geneclean (Bio 101 Inc., Vista, CA, USA). Transgenic mice In situ hybridizations were performed using 35S-labeled were generated by pronuclear injection of the puri®ed riboprobes as described in Fromm and Overbeek (1996). A fragments into one-cell-stage inbred FVB/N embryos by PstI/XbaI fragment of E1A (Figure 1a) was subcloned into standard techniques (Taketo et al., 1991). pBluescript KS+/7, and used to generate an E1A-speci®c riboprobe. A cDNA fragment of b-crystallin (bB1 (*750 bp) was generated by RT ± PCR using an upstream sense primer Screening of mice and histology (5'-CCAATACAGCACCAGGAACC), and a downstream PCR screening was performed as described (Fromm et al., antisense primer (5'-CCGGGAGCTTGTCTCACTTG), and 1994). Embryonic heads at E15.5 were ®xed in 10% formalin, then cloned into pBluescript. A cDNA fragment of g- paran embedded, cut as 5-mm-thick sections, and stained crystallin (g-D1) (*550 bp) was generated by RT ± PCR with hematoxylin and eosin (H&E) by standard techniques. using sense primer (5'-CAGCCATGGGGAAGATCACC)

Oncogene E1A inhibits crystallin expression by binding CBP/p300 Q Chen et al 1036 the same tissue section (counterstained by hematoxylin) using Adobe Photoshop software.

Immunohistochemistry for lens proteins and BrdU Tissue sections were pre-treated as described previously (Robinson et al., 1995). For detection of b- and g-crystallin proteins, rabbit antisera (from Sam Zigler) were used. The biotin-labeled secondary antibody was detected using strep- tavidin conjugated peroxidase, and peroxidase was visualized with diaminobenzidine (DAB)-H2O2 (Vector Laboratories, Kit SK-4100). All slides were incubated in DAB solution for 1 min. Sections were counterstained with hematoxylin. For detection of DNA replication, pregnant females were injected Figure 8 Postulated roles for CBP/p300 in the lens. Our with 5-bromo-2'-deoxyuridine (BrdU) (Sigma), sacri®ced 1 h transgenic studies suggest that induction of late dierentiation later, and embryos were analysed for BrdU incorporation by genes including b- and g-crystallins is lost when CBP and p300 are immunohistochemistry as described (Fromm et al., 1994). For inactivated by E1A (12S and delCR2). Inactivation of Rb by 12S quanti®cation, the number of BrdU positive nuclei in the or delN is predicted to cause release of E2F transcription factors. posterior of the lens was counted and compared to the total Free E2F can induce the cells to enter the cell cycle by upregulating expression of cyclin family members (A2, B1, D3, number of nuclei in the same region as determined by E). This results in inappropriate ®ber cell proliferation. By an hematoxylin staining. unknown mechanism, p53 is activated or stabilized by the inappropriate proliferation. If CBP and p300 are not inactivated, Detection of apoptosis p53 can induce p21. However, when CBP and p300 are inactivated by 12S or delCR2, induction of p21 does not occur DNA fragmentation was detected by TUNEL assay using an in situ apoptosis detection kit according to the manufacturer's instructions (apoTACSTM: In Situ Apoptosis Detection Kit, Trevigen, Inc.). Sections were then counterstained with Methyl Green or hematoxylin. For quanti®cation, the number of apoptotic nuclei was counted and compared to and antisense primer (5'-GCGTCCAAATGGAGAAAGGG), the total number of nuclei as determined by Methyl Green or and also cloned into pBluescript. The cDNAs for other hematoxylin staining. mouse genes were obtained as follows: p21 and p57 from S Elledge (Baylor College of Medicine), c-Maf (maf-2) from Lixing Reneker (Mason Eye Institute, University of Mis- Acknowledgments souri), E2F1 from Peggy Farnham (University of Wisconsin), We thank Greg Barsh, Brian Ring and Michael Schneider cyclin E from Julie A Deloia (Magee-Women's Research for helpful comments on the manuscript; Gaby Schuster Institute), cyclins A2 and B1 from Debra Wolgemuth for generating the transgenic mice; Long Vien, Barbara (Columbia University), p53 from Gigi Lozano (M.D. Harris and Rachel Charbonneau for excellent technical Anderson Cancer Center), and Bax from Stanley Korsmeyer assistance. This research was supported by NRSA fellow- (Washington University). Hybridization signals in Figures 3, ship EY-06708 (JD Ash), NIH grants EY-10448, EY- 4, 6 and 7 were initially collected as dark ®eld images, then 10803, EY-11348 (PA Overbeek); and a grant from the pseudocolor red, and superimposed on bright ®eld images of National Cancer Institute of Canada (P Branton).

References

Bagchi S, Raychaudhuri P and Nevins JR. (1990). Cell, 62, Jones N. (1995). Curr. Top. Microbiol. Immunol., 199, 59 ± 659 ± 669. 80. Bannister AJ and Kouzarides T. (1995). EMBO J., 14, 4758 ± Lill NL, Grossman SR, Ginsberg D, DeCaprio J and 4762. Livingston DM. (1997). Nature 387, 823 ± 827. Barbeau D, Charbonneau R, Whalen SG, Bayley ST and McAvoy JW. (1978). J. Embryol. Exp. Morphol., 44, 149 ± Branton PE. (1994). Oncogene, 9, 359 ± 373. 165. Chrivia JC, Kwok RP, Lamb N, Hagiwara M, Montminy Moran E. (1993). Curr. Opin. Genet. Dev., 3, 63 ± 70. MR and Goodman RH. (1993). Nature, 365, 855 ± 859. Morgenbesser SD, Williams BO, Jacks T and DePinho RA. Eckner R, Ewen ME, Newsome D, Gerdes M, DeCaprio JA, (1994). Nature, 371, 72 ± 74. Lawrence JB and Livingston DM. (1994). Genes Dev., 8, Mymryk JS, Lee RW and Bayley ST. (1992). Mol. Cell. Biol., 869 ± 884. 3, 1107 ± 1115. Eckner R, Yao TP, Oldread E and Livingston DM. (1996). Nishiguchi S, Wood H, Kondoh H, Lovell-Badge R and Genes Dev., 10, 2478 ± 2490. Episkopou V. (1998). Genes Dev., 12, 776 ± 781. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Ogryzko VV, Schiltz RL, Russanova V, Howard BH and Trent JM, Lin D, Mercer WE, Kinzler KW and Vogelstein Nakatani Y. (1996). Cell, 87, 953 ± 959. B. (1993). Cell, 75, 817 ± 825. Piatigorsky J. (1981). Dierentiation, 19, 134 ± 153. Fromm L, Shawlot W, Gunning K, Butel JS and Overbeek Reynaud EG, Pelpel K, Guillier M, Leibovitch MP and PA. (1994). Mol. Cell. Biol., 14, 6743 ± 6754. Leibovitch SA. (1999). Mol. Cell. Biol., 19, 7621 ± 7629. Fromm L and Overbeek PA. (1996). Oncogene, 12, 69 ± 75. Ring BZ, Cordes SP, Overbeek PA and Barsh GS. (2000). Gu W and Roeder RG. (1997). Cell, 90, 595 ± 606. Development, 127, 307 ± 317.

Oncogene E1A inhibits crystallin expression by binding CBP/p300 Q Chen et al 1037 Robinson ML, Overbeek PA, Verran DJ, Grizzle WE, Treton JA, Jacquemin E, Courtois Y and Jeanny JC. (1991). Stockard CR, Friesel R, Maciag T and Thompson JA. Dierentiation, 47, 143 ± 147. (1995). Development, 121, 505 ± 514. Trouche D and Kouzarides T. (1996). Proc. Natl. Acad. Sci. Robinson ML and Overbeek PA. (1996). Invest. Ophthalmol. USA, 93, 1439 ± 1442. Vis. Sci., 37, 2276 ± 2284. Wigle JT, Chowdhury K, Gruss P and Oliver G. (1999). Nat. Smith ML, Kontny HU, Bortnick R and Fornace Jr AJ. Genet., 21, 318 ± 322. (1997). Exp. Cell Res., 230, 61 ± 68. Yao TP, Oh SP, Fuchs M, Zhou ND, Ch'ng LE, Newsome Somasundaram K and El-Deiry WS. (1997). Oncogene, 14, D, Bronson RT, Li E, Livingston DM and Eckner R. 1047 ± 1057. (1998). Cell, 93, 361 ± 372. Taketo M, Schroeder AC, Mobraaten LE, Gunning KB, Zhang P, Wong C, DePinho RA, Harper JW and Elledge SJ. Hanten G, Fox RR, Roderick TH, Stewart CL, Lilly F, (1998). Genes Dev., 12, 3162 ± 3167. Hansen CT et al. (1991). Proc. Natl. Acad. Sci. USA, 88, 2065 ± 2069.

Oncogene