The spatio-temporal expression of three crystallin genes (αA, βB1 and γ) in the developing and regenerating lenses of newt was compared by in situ hybridization in lens differentiation in normal development with during regeneration. In normal development, all crystallin transcripts were first detected at the same stage in the posterior region of the lens vesicle (McDevitt's lens development stage V) and continued during lens fiber differentiation of the posterior cells into the primary lens fiber cell differentiation (McDevitt's lens development stage VII–VIII). At later stages, the expression of the three genes was restricted to the secondary lens fibers and gradually became undetectable in primary lens fibers (McDevitt's lens development stage X). The signal for γ-crystallin was never detected in lens epithelium at any stage, whereas signals for αA- and βB1-crystallin were detected in the lens epithelium at the stage when the primary lens fiber mass was formed. During lens regeneration, signals for the three crystallins were first detected at the same stage at the ventral margin of a regenerating lens vesicle (Sato's lens regeneration stage IV). The expression patterns of three crystallin genes were similar to those in normal development (Sato's lens regeneration stage V–X). The expression pattern of the crystallin genes in normal lens development fundamentally resembles that during lens regeneration, suggesting the absence of unique expression programs of crystallin genes for lens regeneration not found in ontogeny.
In the newt, the lens can be regenerated from the pigmented epithelial cells of the dorsal marginal iris through their transdifferentiation (Reyer 1954; Eguchi 1979). The iris is derived from the neuroectoderm, whereas lens develops from the surface ectoderm through the inductive influence of the non-lenticular tissues in normal development. Nevertheless, the regenerated lens, when completed, is morphologically indistinguishable from the authentically developed lens. Whether or not lens regeneration is controlled by the same program of gene expression as in embryonic lens development without the presence of any special program is a fundamental question that must be answered to understand the mechanism involved in lens regeneration from the dorsal iris in the newt.
Using immunohistochemical techniques, McDevitt and Brahma (1981, 1982, 1990) reported that α-crystallin appeared in the epithelium of regenerating lenses of newts, whereas it was not detected in the lens epithelium during normal lens development, and that the timing of α-crystallin expression was different between normal development and regeneration of lens. These findings support the presence of a different mechanism of crystallin gene expression during lens regeneration from normal development, although McDevitt and Brahma did not study expression of genes encoding major lens proteins, such as α-, β- and γ-crystallin.
To examine the above possibility at the level of expression of corresponding genes, we investigated spatio-temporal expression of α-, β- and γ-crystallin genes in normal lens development and lens regeneration by means of an in situ hybridization technique. Our results suggest that lens regeneration is similar to normal lens development in terms of the expression of crystallin genes.
Materials and Methods
Adult newts (Cynops pyrrhogaster) were collected in Tagarasu district, Obama, Fukui Prefecture, and maintained at 10°C. Female newts were injected with 200 units gonadotropin (Teikokuzouki, Tokyo, Japan) every other day for 1 week. The fertilized eggs and embryos obtained were maintained at 20°C. Removal of lenses from adult newt eyes was performed as follows. After the newts were anesthetized with 0.001% 3-aminobenzonicacidethylester (Sigma, St Louis, MO, USA) in fresh water, a nano-temporal incision was made in the cornea with a razor blade and the lens was removed carefully through the incision with fine forceps. Thereafter, newts were kept at 20°C for up to 60 days. The staging of newt lens regeneration was determined according to Sato (1940): stages III and IV, completion of depigmentation of the dorsal margin of the iris pigmented epithelial cells and formation of the monolayered lens vesicle; stages V–VII, elongation of the posterior cells of the lens vesicle and formation of primary lens fibers; stage VIII, initiation of the secondary lens fiber cells that will eventually surround the primary lens fiber mass. To determine the stages of newt lens development, we followed the stages of McDevitt et al. (McDevitt et al. 1969; McDevitt & Brahma 1981), which can be summarized as follows: at stages IV and V, the lens placode forms a sphere from which the lens vesicle develops, but thickening of the posterior region of the lens vesicle is not observed. At stages VI–IX, cells in the posterior wall of the lens vesicle elongate and become primary lens fiber cells and, beginning at stage X, the secondary lens fiber cells are observed surrounding the primary lens fiber mass.
cDNA cloning and sequencing
To isolate RNA, approximately 1000 regenerated lenses were collected from adult newts. The lenses were washed and homogenized in a solution of 0.1 m Tris-HCl, pH 9.0, and 1% sodium dodecyl sulfate (SDS). Whole nucleic acids were extracted using phenol or phenol–chloroform saturated with 0.1 m Tris-HCl, pH 9.0, and RNA was purified by centrifugation on 5.7 m CsCl. A λgt11 library of approximately 1 × 105 independent clones was constructed from poly(A)+ RNA of the regenerated lenses (Huynh et al. 1985). To isolate αA- and γ-crystallin cDNA, anti-αA-crystallin monoclonal antibody and anti-γ-crystallin antiserum, the specificity of which was confirmed (Sawada et al. 1993), were used to screen a cDNA library (Huynh et al. 1985). Chicken βB1-crystallin cDNA probes isolated in our laboratories were hybridized to obtain β-crystallin cDNA from the library. The hybridization was performed at 60°C and followed by washing once with 2× SSC at room temperature and twice with 0.1× SSC at 50°C. Positive clones were subcloned into the EcoRI site of the plasmid Bluescript KS+ (Stratagene, La Jolla, CA, USA) and sequenced by the dideoxy terminator method using an ABI autosequencer (model 373; Applied Biosystems, Foster City, CA, USA). The nucleotide sequence of αA-, βB1- and γ-crystallins in the present paper will appear in the DDBJ, EMBL and GenBank nucleotide sequence databases with the following accession numbers: D86299, D86300 and D86301, respectively.
In situ hybridization
In situ hybridization was performed according to the method of Iio et al. (1994) with some modification. Briefly, sense and antisense RNA probes for αA-, βB1- and γ-crystallins were synthesized using a DIG RNA labeling kit (Boehringer Mannheim, Mannheim, Germany). Regenerated lenses were taken together with iris and cornea from adult operated eyes and fixed with 4% paraformaldehyde in 80% Ca2+ and Mg2+-free phosphate-buffered saline (PBS). Embryos were fixed with 4% paraformaldehyde in 80% PBS. After fixation, samples were dehydrated in an ethanol series and embedded in paraffin. Then, 6 µm sections were cut and placed on 3-amino-propyltritethoxysilane (Sigma)-coated slide glasses and hybridized with riboprobes for newt αA-, βB1- and γ-crystallin. We used single individuals to make a specimen for three crystallin probes. The hybridization was performed at 50°C for 16 h; then, specimens were washed three times with 0.1× SSC at 50°C and treated with RNase A. Positive signals were detected using a DIG nucleic acid detection kit (Boehringer Mannheim).
Expression of crystallin genes in embryonic newt lens
A signal for αA-crystallin was first detected at stage V in the posterior region of the lens vesicle (Fig. 1A). At stage VII, immediately prior to differentiation of the posterior cells of the lens vesicle into primary lens fibers, the αA-crystallin signal was detected at the elongated posterior cells of the vesicle (Fig. 1B). After the lens cavity was closed by elongation of the primary lens fiber cells, the signal still remained in all the primary lens fiber cells at stages VIII (Fig. 1C). At stage X, newly formed secondary lens fiber cells and epithelial cells showed the signal, whereas the signal in the primary lens fiber masses gradually disappeared from the center toward the periphery (Fig. 1D).
The signal using the βB1-crystallin probe was first detected in the same region and at the same stage as the αA-crystallin signal (Fig. 2A). A few cells in the anterior lens epithelium began to show the βB1-crystallin signal (stage VII; Fig. 2B). At stage VIII, after the primary lens fiber mass was formed, the βB1-crystallin signal was detected in the entire lens, including the primary lens fiber mass and the lens epithelium (Fig. 2C). At stage X, the signal in the primary lens fiber mass became weaker, as in the case of the αA-crystallin signals (Fig. 2D).
The pattern of the γ-crystallin signal was almost the same as that of the αA-crystallin transcripts, but the γ-crystallin signal was only expressed in primary and secondary lens fiber cells (Fig. 3A–D).
Expression of crystallin genes in regenerating newt lens
At stage IV, the αA-crystallin signal was first detected at the ventral tip of the lens vesicle (Fig. 1E). After the posterior cells of the lens vesicle began to elongate and formed primary lens fiber cells (stage VI), the signal was detected not only in the primary lens fiber cells, but also in a few cells in the ventral epithelium of the lens vesicle (Fig. 1F). When secondary lens fibers began to surround the primary lens fiber mass at stage VIII, a signal was detected in newly formed secondary lens fiber cells. After this stage, signals in the primary lens fiber mass gradually disappeared. The area of lens epithelium in which the αA-crystallin gene was expressed extended from ventral toward anterior, but no cells in the lens stalk connecting the regenerating lens and pigmented dorsal iris showed any signals at stage VIII (Fig. 1G). At stage IX, signals were restricted in young secondary lens fiber cells and epithelial cells located in the equatorial region (Fig. 1H).
The signal for βB1-crystallin was first detected at the same stage and in the same region as αA-crystallin transcripts, but the expression area was larger than for αA-crystallin (stage IV; Fig. 2E). At stage VI, although the βB1-crystallin signal was mainly detected in the primary lens fiber cells, a signal was also detected in a few cells of the anterior and ventral parts of the lens epithelium (Fig. 2F). After formation of the primary lens fiber mass at stage VIII, both lens epithelial cells and young secondary lens fiber cells showed signals. Cells in the lens stalk showed a positive signal at this stage (Fig. 2G). At stage IX, the signal in the primary lens fibers became gradually weaker, but the entire lens epithelium and young secondary lens fiber cells were positive (Fig. 2H).
The signal for γ-crystallin was first detected at the same stage and in the same region as for αA-crystallin (stage IV; Fig. 3E). As lens regeneration progressed, the γ-crystallin signal was detected only in primary lens fiber cells (Fig. 3F,G). When secondary lens fiber formation occurred, the γ-crystallin signals in the primary lens fiber cell masses became gradually weaker, as for αA-crystallin (Fig. 3H).
McDevitt and Brahma (1981, 1982, 1990) investigated α-, β- and γ-crystallin expression in normal development and regeneration of lenses using antibodies. They showed that three major crystallins are expressed in the order of β-, γ- and α- during normal lens development but, during lens regeneration, β- and γ-crystallins appeared simultaneously at the same stage, whereas α-crystallin was expressed later. In the present study, signals for αA-, βB1- and γ-crystallins were all detected at the same stage of lens vesicle formation in both development and regeneration. McDevitt and Brahma reported a spatial difference in α-crystallin expression between normal lens development and lens regeneration (McDevitt & Brahma 1981, 1982, 1990). In normal lens development, α-crystallin was only detected in lens fiber regions, whereas in lens regeneration it was expressed additionally in lens epithelium. Our in situ hybridization results regarding αA-crystallin RNA expression were different from the observations of McDevitt and Brahma. In the present study, the αA-crystallin signal was detected in lens epithelium in normal lens development, which was not found in the previous studies. This difference may be due to the different rates at which the transcripts are translated into the corresponding proteins. Alternatively, the populations hybridized with our probes may differ markedly from those recognized by the antibodies used in the previous studies (McDevitt & Brahma 1981, 1982, 1990).
In newts, despite differing origins of the lens in normal lens development and regeneration, the expression pattern of the three crystallin genes was similar in the two processes. These results suggest that, as far as the major crystallin genes are concerned, the expression program in lens regeneration is a simple repeat of the course of normal lens development in the newt.
In recent years, Mizuno et al. investigated αA-, βB1- and γ-crystallin gene expression in Xenopus by in situ hybridization (Mizuno et al. 1999). Comparison of the appearance of these transcripts in embryonic lens development and in lens regeneration indicated appreciable differences. In embryonic lens, all three crystalline genes were already expressed at the lens placode stage, before formation of lens rudiment. In contrast, during the regeneration process, expression of the γ-crystallin gene is delayed relative to αA- and βB1-crystallins: αA- and βB1-crystallin expression was first detected in lens vesicles just before lens fiber formation, whereas the γ-crystallin signal became detectable only during lens fiber development. It is thus interesting to note that the order of activation of crystallin genes resembles embryonic lens development more in Cynops than in Xenopus, although in Xenopus normal lens and regenerating lens arise from ectoderm.
We are grateful to Dr T. S. Okada (BRH, Takatuki, Japan) for his critical reading of the manuscript, helpful comments and suggestions. We also thank Dr H. Kondoh (Osaka University, Suita, Japan) for his invaluable advice and kind help, and Dr R. Kodama, Mr T. S. Yamamoto, Ms. C. Takagi and Ms. H. Urai (National Institute for Basic Biology, Okazaki, Japan) for their kind assistance. This study was supported by Grants-in-Aid from Japan's Ministry of Education, Science, sports and Culture (to GE and MM).