- Email: [email protected]
Cellular heterogeneity in the expression of the δ-crystallin gene in non-lens tissues
DEVELOPMENTAL
BIOLOGY
112,94-99
(1985)
Cellular Heterogeneity in the Expression of the b-Crystallin Gene in Non-lens Tissues J-C. JEANNY,* D.J. BOWER, L.H. Department
ERRINGTON,S.MORRIS,
AND
R.M. CLAYTON'
of Animal Genetics, University of Edinburgh, West Mains Road, Edinburgh EH9 UN, Scotland, and *Unite de Recherches Gerontologiques, l%lB, 29, Rue Wilhem, 75016 Paris, France Received October 29, 1984; accepted in revised form May 16, 1985
RNA transcripts of the Scrystallin genes, which code for the major chicken lens protein, have been detected at low levels in many non-lens tissues. Here it is demonstrated by in situ hybridisation that these transcripts are concentrated at a high level in small, infrequent clusters of cells in many non-lens tissues. While the nuclei of these cells are very heavily labelled, there is only light labelling of the cytoplasm. The unlabelled cells surrounding the labelled clusters are of similar morphology and staining properties as the labelled cells, and all have the characteristic morphology of cells of the embryonic tissue used. With the exception of neural retina, it is not yet known whether the labelled clusters are found in specific locations in the tissues, or whether they arise at random. o 1985 Academic POW, IX. INTRODUCTION
&Crystallin comprises about 50% of the total protein of the adult chick lens and 70-80% of the 15-day embryo lens (Piatigorsky et al., 1976). Although they have been generally regarded as lens-specific proteins, crystallins have been detected at low levels in several embryonic and post-hatch tissues, including iris, retina, epiphysis, and adenohypophysis (Maisel and Harmison, 1963; Clayton et aL, 1968; Clayton, 1978; Bours and van Doorenmaalen, 1972; Barabanov, 1977; Watanabe, 1984; Watanabe et al, 1983; Ueda et ak, 1983). All these tissues are derived from diencephalon or head ectoderm, in parts of the head where these two embryonic regions became contiguous, with very little separating layer of mesenthyme. They are all able to dedifferentiate and then redifferentiate (transdifferentiate) to lens cells (reviews: Clayton, 1982; Okada, 1983). Crystallin antigenicity at much lower, trace levels are also detectable in several other tissues which have not been reported to transdifferentiate (Clayton et al., 1968). There are two Gcrystallin genes in the chicken genome (Bhat and Piatigorsky, 1979) and they are arranged in tandem on 20 kb of DNA of the same chromosome (Yasuda et al, 1982). While it was concluded that d-crystallin RNA transcripts are present at stages earlier than 3.5 days (Piatigorsky et ab, 1976), fluorescent antibodies have shown that not all cells in the lens vesicle contain Scrystallin protein (Zwaan and Ikeda, 1968), and it is possible that not all of these cells are yet expressing IS’ To whom correspondence
should be addressed.
001%1606185 $3.00 Copyright All rights
0 1985 by Academic Press. Inc. of reproduction in any form reserved.
94
crystallin RNA. However, by 3.5 days all cells of the embryo lens contain b-crystallin RNA (Bower et ah, 1983a); intermediate levels of crystallin RNA have been detected in neural retina (NR) and retinal pigmented epithelium (Jackson et ab, 1978) which have the capacity to transdifferentiate. More recently, Scrystallin transcripts have been found in all tissues known to be able to transdifferentiate (Agata et aZ., 1983; Bower et al., 1983a). In early embryos, in situ hybridisation shows that &crystallin RNA in NR is localised in the nuclei of cells on the vitreal border of the retina (Bower et aL, 1983a). Much smaller amounts of d-crystallin RNA have also been detected in developing chick heart, lung, kidney, and liver (Bower et aZ.,1983b), tissues which have not been shown to transdifferentiate. Previously it was not possible to say whether all the cells in these tissues were transcribing d-crystallin RNA at a low level or whether a subset of cells was transcribing at a high level, although a comparison of the rate of increase of crystallin RNA during transdifferentiation of neural and pigmented retina cells led us to suggest that the cells might be heterogeneous in expression (Thomson et al., 1981) and that transdifferentiation potential depended both on the number of cells expressing crystallin RNA in a tissue and on the levels at which it was expressed per cell. Here we show that most or all of d-crystallin transcription in 4- to 6-day embryo chick NR, epiphysis, adenohypophysis, otic vesicle, and heart takes place in clusters of cells surrounded by cells of similar morphology and staining properties which are not transcribing at a detectable level.
JEANNY
MATERIALS
AND
ET AL.
&Crystallin
METHODS
Embryos of the ‘Shaver’ strain were used. This is a white leghorn-derived line inbred for 14 years at the Poultry Research Centre, Midlothian. The probe, pM56, is a plasmid constructed by inserting 600 bp of &crystallin double-stranded cDNA into the Pat1 site of pBR322 (Bower et ah, 1982). The inserted sequence hybridises to the protein-coding sequences from the 5’ end of both b-crystallin genes, plus a few nucleotides of 5’ noncoding sequence. It contains no repetitive sequence and hybridises exclusively to the two fi-crystallin genes, under conditions of moderate or high stringency. It does not show any tendency to bind weakly to ribosomal RNA or DNA (Bower et ah, 1983b). The probes, pM56 and pBR322, were labelled with 3H to a specific activity of 3 X lo6 dpm by nick translation (Maniatis et al, 1975), and separated from nucleotides by chromatography over Sephadex G-50 (Pharmacia). After ethanol precipitation, probes were resuspended at 2 X lo4 dpm/yl in 2X SSC, 38% formamide and denatured by incubating at 100°C for 5 min. Embryonic tissues were rapidly dissected out and immediately squashed on sterile slides (Sandritter et al., 1966). Fixation and hybridisation were as described in Maitland et al, (1981). Washing, including a thermal wash at 55°C in 2 changes of 0.1X SSC (1X SSC = 0.15 M NaCl, 0.015 M Na citrate), autoradiography, and staining were as described previously (Bower et al., 1983a). The speed required for dissection to prevent RNA degradation precluded the possibility of obtaining 100% pure tissue from epiphysis and adenohypophysis. Thus, it was not possible to assess accurately the percentage of labelled cells in these tissues. A parallel series of squashes on slides was pretreated with RNase as follows: after fixation and dehydration, 50 ~1 of 100 pg/ml RNase A (Sigma) in 2X SSC was placed on each squash, covered with a glass coverslip, and incubated at 25°C for 30 min. Slides were washed twice for 15 min in 2~ SSC and then dehydrated through the alcohol series as before. These squashes were then hybridised to 3H-labelled pM56 as described for untreated squashes. DNase treatment of squashes. Another parallel series of squashes on slides was pretreated with DNase as follows: after fixation and dehydration, 50 ~1 of 100 pg/ml RNase-free DNase in 10 mM Tris, pH 7.4, was placed on each squash, covered with a glass coverstop, and incubated at 37°C for 30 min. Slides were washed twice for 15 min in 2~ SSC and then dehydrated through the alcohol series as before. These squashes were then hybridised to 3H-labelled pM56 as described for untreated squashes.
in
Non-lens
95
Tissues RESULTS
3H-Labelled pM56 was hybridised to squashes of tissue from 4- to 6-day lens, neural retina, retina pigment epithelium (not shown), adenohypophysis, epiphysis, otic vesicle, and heart. Parallel squashes were hybridised to 3H-labelled pBR322, the plasmid moiety of pM56. After a 6-week exposure, lens cells were all uniformly labelled by pM56 over cytoplasm and nuclei (data not shown). All non-lens tissues contained clusters of heavily labelled cells, with grains predominantly over the nuclei, surrounded by large areas of unlabelled cells with morphology and staining properties similar to the labelled cells (Fig. 1). Control squashes of all tissue types hybridised to 3H-labelled pBR322 contained no labelled cells (5-day lens only shown, Fig. 2D). Control squashes which were pretreated with RNase before hybridisation were negative (Fig. 2C), but DNase pretreatment did not prevent hybridisation (Fig. 2B). While labelled and unlabelled cells within a tissue were similar, they were morphologically distinct from cells in other tissues. Thus in each non-lens tissue examined, cells characteristic of that tissue were divided into two classes, a small class transcribing large amounts of &crystallin RNA, and a large class which did not transcribe the 6-crystallin genes, at least at a level detectable by in situ hybridisation. DISCUSSION
Since the plasmid vector pBR322 gave no hybridisation to squashes, and pM56 hybridises with high specificity to the two d-crystallin genes only, the heavy nuclear labelling of clusters of cells is due not to nonspecific binding, but to the presence of very large amounts of &crystallin RNA in the nuclei of these clusters. This is confirmed by the abolition of hybridisation by pretreatment with RNase. The absence of silver grains over cells with morphology similar to the labelled cells also indicates that labelling is not due to nonspecific sticking to some types of cells. We cannot precisely quantitate the number of transcripts, since different preparations retain accessible RNA to different extents, but with the fixation methods used here and the specific activity of the probe, several thousands of transcripts would be required to give the level of labelling observed (Hafen et ah, 1983; Angerer and Angerer, 1981). Since there are only 4-8 copies of the a-crystallin genes per diploid cell, and in any case only a few cells are labelled, we can rule out the possibility that the probe was hybridising to DNA. Furthermore the hybridisation is unaffected by pretreatment with DNase. The lower levels of labelling of cytoplasm might be due to some spreading from the nuclei during squashing, but data from Northern transfers (Bower et al, 1983b)
96
DEVELOPMENTAL BIOLOGY
VOLUME 112, 1985
FIG. 1. Embryonic tissues were dissected out, squashed, fixed, hybridised to ‘H-labelled pM56, dipped in Ilford K-2 nuclear emulsion diluted 1:l with HzO, dried, and exposed for 6 weeks at 4°C in a light-proof box. The original magnifications of the photographs are given. (A) 3.5- to 4-day neural retina, x400; (B) 5-day heart, X400; (C) 3.5- to 4-day epiphysis, X1000; (D) g-day adenohypophysis, X1000; (E) 3.5- to $-day otic vesicle, X1000; (F) g-day neural retina, X400.
indicate that while most of the RNA in non-lens tissues hybridising to pM56 (Scrystallin cDNA clone) was larger than the major d-crystallin mRNA, now known from sequence data to be 15’74nucleotides plus poly(A) (Yasuda et al., 1984; Nicker-son and Piatigorsky, 1984), a small percentage of the hybridising RNA in non-lens
tissues comigrated with the major lens Scrystallin mRNA. In addition, the size classes of RNA which hybridised to pM56 were different between nuclear and cytoplasmic RNA from non-lens tissues (Bower et ah, 1983b), suggesting that the cytoplasmic material detected was genuine and was not due to contamination
JEANNY ET AL.
6Gystallin
in Nm-lens
Tissues
FIG. 2. Embryonic 3.5-day lenses were treated and hybridised as for Fig. 1. (A) Hybridised to 3H-labelled pM56 as for Fig. 1, X1000. (B) Squash pretreated with DNase before hybridisation, X400. Hybridisation has not been prevented. (C) Squash pretreated with RNase before hybridisation, X400. (D) Squash hybridised to 3H-labelled pMB322, the plasmid moiety of pM56, X400.
by nuclear RNA. This implies that the light labelling detected over cytoplasm of cells with heavily labelled nuclei in Figs. lA, C, and D may be due to b-crystallin RNA being transported to the cytoplasm. Significant cytoplasmic label was not seen in other tissues (Figs. 1B and E) which are not known to have transdifferentiation potential. Low levels of crystallin antigenicity have been detected in all tissues of the eye (Maisel and Harmison, 1963; Clayton et ab, 1968; Bours and van Doorenmaalen, 1972) and at very low, or trace levels in several extraocular tissues including brain, liver, kidney, muscle, and skin (Clayton et al., 1968). More recently, immunohistology has shown cells with &crystallin antigenicity in adenohypophysis brain (Barabanov, 19’77, 1982) and epiphysis (Ueda et aZ., 1983; Watanabe et al., 1983). The detection of crystallin specificity does not, however, necessarily imply the presence of a completed polypeptide (Clayton et ah, 1968; Clayton, 1978), but these data suggest that some translation may be taking place on these templates.
Schibler et al. (1983) have found a similar pattern of expression in the a-amylase la gene, which is expressed in several tissues, including the brain, where RNA transcripts are confined to the nucleus, although it is as abundant as in liver, where the RNA is translated. They have not yet established whether all brain cells transcribe the gene, or only a subset, as we find here for the &crystallin genes. Mouse antitrypsin and antichymotrypsin genes, expressed at a high level in liver, are expressed at a low level in other tissues (Shaw et ah, 1983). Again, it is not known if expression is at a high level in a few cells, as we find for &crystallin genes, or at a low level throughout. We do not know whether the pattern of &crystallin gene transcription is the same in expressing non-lens cells as lens. It may be that of the two crystallin genes, a different one is transcribed in lens compared with other tissues. However, while &crystallin protein detected in NR is indistinguishable from that in the lens, the 6crystallin protein detected in the adenohypophysis differs electrophoretically from lens &crystallin (de Pom-
98
DEVELOPMENTAL BIOLOGY
erai and Clayton, 1978; Barabanov, 1977). If different genes are transcribed in different tissues, this would parallel the case of the a2u-globulin genes (Laperche et ah, 1983) and the MUP genes (Shaw et aZ.,1983). Another possibility is that the promoter used is different, as with the a-amylase la gene (Schibler et al., 1983), the ovalbumin gene (Malek et aZ.,1981), and the lysozyme gene (Grez et ah, 1981). Our present data show that the predominantly highmolecular-weight d-crystallin RNA transcripts previously detected in tissues unrelated to the eye (Bower et al., 1983a) and in eye-associated tissues with transdifferentiation capacity are due to high-level transcription in a few cells rather than low level in all. Heterogeneity with respect to methylation of the Scrystallin genes has also been observed in g-day NR (Cooper et ah, 1983). In NR a specifically located subpopulation appears to be involved in d-crystallin transcription (Bower et ab, 1983b), but in other tissues it is not yet known whether this is the case; no morphological or staining difference has been found between labelled and unlabelled cells, suggesting that they may all be cells of the same subtype. The clustering of labelled cells in the tissues examined suggests either that cells in a cluster are clonally related or that they are in a special location in which an environmental factor induces Scrystallin gene expression. We are currently investigating these possibilities. Higher levels of RNA detected in NR compared with heart are probably due to a higher proportion of cells expressing the genes in NR, although we do not yet have precise figures for the same embryonic stage in both tissues. On serial sections of the central part of the 3.5 day eye, 15% of NR cells were labelled (unpublished data), and in squashes of the whole &day heart, 0.1% of cells were labelled. The number of labelled cells in a cluster was higher for neural retina than for heart. However, we do not have figures for the entire, total 3.5day NR, and it is possible that there is some difference between different parts of the retina, since the pattern of labelling is not random, but localised to the vitreal border. In addition, there may be some change between 3.5 and 5 days in percentage of cells labelled. There appear to be cells, sometimes within a single tissue, exhibiting several different transcriptional states of the Scrystallin genes: those in which it would appear that these crystallin sequences are not transcribed at detectable levels, those in which they are transcribed at high levels and rapidly processed, as in embryo lens, and those in which there is appreciable transcription as in the non-lenticular tissues examined here, some of which appear also to show some degree of processing. The level of processing may be higher in retina than in otic vesicle, and we present elsewhere evidence that retina cells have
VOLUME 112, 1985
latent processing capacity (Clayton et aL, in press; Bower et al., submitted for publication). We are now looking for expression of other crystallin genes in these tissues, and investigating the fate of crystallin gene-expressing clusters in development. This work was supported by the Medical Research Council. J-C. Jeanny was supported by a Royal Society Visiting Research Fellowship. We are grateful to the Poultry Research Centre, Roslin, Midlothian, for fertile eggs.
REFERENCES AGATA, K., YASUDA, K., and OKADA, T. S. (1983). Gene coding for a lens-specific protein, 8-crystallin, is transcribed in non-lens tissues of chicken embryos. Dev. Biol. 100,222-226. ANGERER, L. M., and ANGERER, R. C. (1981). Detection of poly A + RNA in sea urchin eggs and embryos by quantitative in situ hyhridisation. Nucleic Acids Res. 9, 2819-2840. BARABANOV, V. M. (1977). Detection of &crystallins in the adenohypophysis of the chick embryo (Russian). Dokl. Akcd. Nauk. SSSR 234,195-198. BARABANOV, V. M. (1982). Extra lenticular localisation of b-crystallin in Japanese quail (Russian). Dokl. Akad. Nauk. SSSR. 262,22-24. BHAT, S. P., and PIATIGORSKY, J. (1979). Molecular cloning and partial characterisation of 6-crystallin cDNA sequence in a bacterial plasmid. Proc. NatZ. Acad. Sci. USA 76,3299-3303. BOURS, J., and VAN DOORENMAALEN, W. J. (1972). The presence of lens antigens in intraocular tissues of the chick eye. Exp. Eye Res. 13, 236-247. BOWER, D. J., ERRINGTON, L. H., WAINWRIGHT, N. R., SIME, C., MORRIS, S., and CLAYTON, R. M. (1982). Cytoplasmic RNA sequences complementary to cloned chick &crystallin cDNA show size heterogeneity. Biochem. J 201,339-344. BOWER, D. J., ERRINGTON, L. H., POLLOCK, B. J., MORRIS, S., and CLAYTON, R. M. (1983a). The pattern of expression of chick &crystallin genes in lens differentiation and in transdifferentiating cultured tissues. EMBO J. 2, 333-338. BOWER, D. J., ERRINGTON, L. H., COOPER,D. N., MORRIS, S., and CLAYTON, R. M. (1983b). Chicken lens &crystallin gene expression and methylation in several non-lens tissues. Nucleic Acids Res. 11,2513-2527. BOWER, D. J., JEANNY, J-C., ERRINGTON, L. H., MORRIS, S., and CLAYTON, R. M. (1985). A quantitative comparison of &crystallin gene expression during development of chick lens and non-lens tissues. Submitted for publication. CLAYTON, R. M. (1978). Divergence and convergence in lens cell differentiation: Regulation of the formation and specific content of the lens fibre cell. In “Stem Cells and Tissue Homeostasis,” (B. I. Lord, C. S. Potten, and R. J. Cole, eds), C.U.P., pp 115-138. CLAYTON, R. M. (1982). The molecular basis for competence, determination and transdifferentiation: A hypothesis. A&J. Exp. Biol. Med. 158,23-38. CLAYTON, R. M., CAMPBELL, I. C., and TRUMAN, D. E. S. (1968). A reexamination of the organ specificity of lens antigens. Ezp. Eve Res. 7, 11-29. CLAYTON, R. M., JEANNY, J-C., BOWER, D. J., and ERRINGTON, L. H. (1985). Extra lenticular crystallins and their role during transdifferentiation to lens. Curr. Top Dev. Bid, in press. COOPER,D. N., ERRINGTON, L. H., and CLAYTON, R. M. (1983). Variation in the DNA methylation pattern of expressed and nonexpressed genes in chicken. DNA 2,129-138.
JEANNY ET AL.
6-Crystallin
GREZ, M., LAND, H., GIESECKE, K., SCHUTZ, G., JUNG, A., and SIPPEL, A. E. (1981). Multiple mRNAs are generated from the chicken lysozyme gene. Cell 25,743-752. HAFEN, E., LEVINE, M., GARBER, R. L., and GEHRING, W. J. (1983) An improved in situ hybridisation method for the detection of cellular RNAs in Drosophila tissue sections and its applications for localising transcripts of the homeotic Antennapedia gene complex. EMBO J. 2, 617-623. JACKSON, J. F., CLAYTON, R. M., WILLIAMSON, R., THOMSON, I., TRUMAN, D. E. S., and DE POMERAI, D. I. (1978). Sequence complexity and tissue distribution of chick lens crystallin mRNAs. Dev. Biol 65, 383-395. LAPERCHE, Y., LYNCH, K. R., DOLAN, K. P., and FEIGELSON, P. (1983). Tissue-specific control of 012~globulin gene expression: Constitutive synthesis in the sub-maxillary gland. Cell 32,453. MAISEL, H., and HARMISON, C. (1963). An immunoembryological study of the chick iris. J. EmbyoL Exp. MorphoL 11,483-491. MAITLAND, N. J., KINROSS, J. H., BUSU~~IL, A., LUDGATE, S. M., SMART, G. E., and JONES, K. W. (1981). Detection of DNA tumour virusspecific RNA sequences in abnormal human cervical biopsies by in situ hybridisation. J. Gen. ViroL 55, 123-137. MALEK, L. T., ESCHENFELDT, W. H., MUNNS, T. W., and RHOADS, R. E. (1981). Heterogeneity of the 5’terminus of lens ovalbumin mesenger ribonucleic acid. Nucleic Acids Res. 9, 1657-1673. MANIATIS, T., JEFFREY, A., and KLEID, D. G. (1975). Nucleotide sequence of the rightward operator of phage lambda. Proc. NatL Acad Sci. USA 72.1184-1188. NICKERSON, J. M., and PIATIGORSKY, J. (1984). Sequence of a complete chicken 6-crystallin cDNA. Proc. NatL Acad. Sci. USA 81,2011-2015. OKADA, T. S. (1983). Cellular and molecular aspects of transdifferentiation of ocular tissues in vitro. Cell &%r. 12, 117-184. PIATIGORSKY, J., BEEBE, D., ZELENKA, P., MILSTONE, L. M., and SHINOHARA, T. (1976). Regulation of &crystallin gene expression during development of the embryonic chick lens. INSERM 60,85-112. DE POMERAI, D. I., PRITCHARD, D. J., and CLAYTON, R. M. (1977). Biochemical and immunological studies of lentoid formation in cultures
in Nor-lens
Tissues
99
of embryonic chick neural retina and day-old chick lens epithelium. Dev. BioL 60,416-427. DE POMERAI, D. I., and CLAYTON, R. M. (1978). Influence of embryonic stage on the transdifferentiation of chick neural retina cells in culture. J. EmbryoL Exp. MorphoL 47, 178-193. SANDRITTER, W., PILNY, V., NOVAKOVA, V., and KIEFER, G. (1966). Zur Problematik der Gewebspraparation fur cytophotometrische Messungen. Histochemie 7, l-7. SCHIBLER, U., HAGENBUCHLE, O., WELLAUER, P. K., and PIITET, A. C. (1983). Two promoters of different strength control the transcription of the mouse alpha-amylase gene Amy-la in the parotid gland and the liver. Cell 33, 501-508. SHAW, P. H., HELD, W. A., and HASTIE, N. D. (1983). The gene family for major urinary proteins: Expression in several secretory tissues of the mouse. Cell 32,755-761. THOMSON, I., YASUDA, K., DE POMERAI, D. I., CLAYTON, R. M., and OKADA, T. S. (1981). The accumulation of lens-specific protein and mRNA in cultures of neural retina from 3.5 day chick embryos. Exp. Cell Res. 135, 445-449. UEDA, G., HIRO, F., TAKEICHI, M., and OKADA, T. S. (1983). Detection of 6-crystallin in adenohypophysis in chick embryos using monoclonal antibodies. Dev. Growth Difler. 25,421. WATANABE, K. (1984). In “Proceedings, 12th NIBB Conf. Molecular Mechanisms of Cell Specialisation in Development.” WATANABE, K., TAMAMAKI, N., and NOJYO, Y. (1983). Probable presence of oculopotent cells in embryonic pineal of quail. Dev. Growth Difer. 25, 421. YASUDA, K., KONDOH, H., OKAZAKI, T. S., OKADA, T. S., and SHIMURA, Y. (1982). Organization and structure of (Y- and &crystallin genes and their expression in cells by microinjection. In “Proceedings, Vth Int. Cong. Eye Res., Eindhoven,” p. 38. YASUDA, K., NAKAJIMA, N., ISOBE, T., OKADA, T. S., and SHIMURA, Y. (1984). The nucleotide sequence of a complete chicken &crystallin cDNA. EMBO J. 3,1397-1402. ZWAAN, J., and IKEDA, A. (1968). Macromolecular events during differentiation of the chick lens. Exp. Eye Res. 7,301-311.
BIOLOGY
112,94-99
(1985)
Cellular Heterogeneity in the Expression of the b-Crystallin Gene in Non-lens Tissues J-C. JEANNY,* D.J. BOWER, L.H. Department
ERRINGTON,S.MORRIS,
AND
R.M. CLAYTON'
of Animal Genetics, University of Edinburgh, West Mains Road, Edinburgh EH9 UN, Scotland, and *Unite de Recherches Gerontologiques, l%lB, 29, Rue Wilhem, 75016 Paris, France Received October 29, 1984; accepted in revised form May 16, 1985
RNA transcripts of the Scrystallin genes, which code for the major chicken lens protein, have been detected at low levels in many non-lens tissues. Here it is demonstrated by in situ hybridisation that these transcripts are concentrated at a high level in small, infrequent clusters of cells in many non-lens tissues. While the nuclei of these cells are very heavily labelled, there is only light labelling of the cytoplasm. The unlabelled cells surrounding the labelled clusters are of similar morphology and staining properties as the labelled cells, and all have the characteristic morphology of cells of the embryonic tissue used. With the exception of neural retina, it is not yet known whether the labelled clusters are found in specific locations in the tissues, or whether they arise at random. o 1985 Academic POW, IX. INTRODUCTION
&Crystallin comprises about 50% of the total protein of the adult chick lens and 70-80% of the 15-day embryo lens (Piatigorsky et al., 1976). Although they have been generally regarded as lens-specific proteins, crystallins have been detected at low levels in several embryonic and post-hatch tissues, including iris, retina, epiphysis, and adenohypophysis (Maisel and Harmison, 1963; Clayton et aL, 1968; Clayton, 1978; Bours and van Doorenmaalen, 1972; Barabanov, 1977; Watanabe, 1984; Watanabe et al, 1983; Ueda et ak, 1983). All these tissues are derived from diencephalon or head ectoderm, in parts of the head where these two embryonic regions became contiguous, with very little separating layer of mesenthyme. They are all able to dedifferentiate and then redifferentiate (transdifferentiate) to lens cells (reviews: Clayton, 1982; Okada, 1983). Crystallin antigenicity at much lower, trace levels are also detectable in several other tissues which have not been reported to transdifferentiate (Clayton et al., 1968). There are two Gcrystallin genes in the chicken genome (Bhat and Piatigorsky, 1979) and they are arranged in tandem on 20 kb of DNA of the same chromosome (Yasuda et al, 1982). While it was concluded that d-crystallin RNA transcripts are present at stages earlier than 3.5 days (Piatigorsky et ab, 1976), fluorescent antibodies have shown that not all cells in the lens vesicle contain Scrystallin protein (Zwaan and Ikeda, 1968), and it is possible that not all of these cells are yet expressing IS’ To whom correspondence
should be addressed.
001%1606185 $3.00 Copyright All rights
0 1985 by Academic Press. Inc. of reproduction in any form reserved.
94
crystallin RNA. However, by 3.5 days all cells of the embryo lens contain b-crystallin RNA (Bower et ah, 1983a); intermediate levels of crystallin RNA have been detected in neural retina (NR) and retinal pigmented epithelium (Jackson et ab, 1978) which have the capacity to transdifferentiate. More recently, Scrystallin transcripts have been found in all tissues known to be able to transdifferentiate (Agata et aZ., 1983; Bower et al., 1983a). In early embryos, in situ hybridisation shows that &crystallin RNA in NR is localised in the nuclei of cells on the vitreal border of the retina (Bower et aL, 1983a). Much smaller amounts of d-crystallin RNA have also been detected in developing chick heart, lung, kidney, and liver (Bower et aZ.,1983b), tissues which have not been shown to transdifferentiate. Previously it was not possible to say whether all the cells in these tissues were transcribing d-crystallin RNA at a low level or whether a subset of cells was transcribing at a high level, although a comparison of the rate of increase of crystallin RNA during transdifferentiation of neural and pigmented retina cells led us to suggest that the cells might be heterogeneous in expression (Thomson et al., 1981) and that transdifferentiation potential depended both on the number of cells expressing crystallin RNA in a tissue and on the levels at which it was expressed per cell. Here we show that most or all of d-crystallin transcription in 4- to 6-day embryo chick NR, epiphysis, adenohypophysis, otic vesicle, and heart takes place in clusters of cells surrounded by cells of similar morphology and staining properties which are not transcribing at a detectable level.
JEANNY
MATERIALS
AND
ET AL.
&Crystallin
METHODS
Embryos of the ‘Shaver’ strain were used. This is a white leghorn-derived line inbred for 14 years at the Poultry Research Centre, Midlothian. The probe, pM56, is a plasmid constructed by inserting 600 bp of &crystallin double-stranded cDNA into the Pat1 site of pBR322 (Bower et ah, 1982). The inserted sequence hybridises to the protein-coding sequences from the 5’ end of both b-crystallin genes, plus a few nucleotides of 5’ noncoding sequence. It contains no repetitive sequence and hybridises exclusively to the two fi-crystallin genes, under conditions of moderate or high stringency. It does not show any tendency to bind weakly to ribosomal RNA or DNA (Bower et ah, 1983b). The probes, pM56 and pBR322, were labelled with 3H to a specific activity of 3 X lo6 dpm by nick translation (Maniatis et al, 1975), and separated from nucleotides by chromatography over Sephadex G-50 (Pharmacia). After ethanol precipitation, probes were resuspended at 2 X lo4 dpm/yl in 2X SSC, 38% formamide and denatured by incubating at 100°C for 5 min. Embryonic tissues were rapidly dissected out and immediately squashed on sterile slides (Sandritter et al., 1966). Fixation and hybridisation were as described in Maitland et al, (1981). Washing, including a thermal wash at 55°C in 2 changes of 0.1X SSC (1X SSC = 0.15 M NaCl, 0.015 M Na citrate), autoradiography, and staining were as described previously (Bower et al., 1983a). The speed required for dissection to prevent RNA degradation precluded the possibility of obtaining 100% pure tissue from epiphysis and adenohypophysis. Thus, it was not possible to assess accurately the percentage of labelled cells in these tissues. A parallel series of squashes on slides was pretreated with RNase as follows: after fixation and dehydration, 50 ~1 of 100 pg/ml RNase A (Sigma) in 2X SSC was placed on each squash, covered with a glass coverslip, and incubated at 25°C for 30 min. Slides were washed twice for 15 min in 2~ SSC and then dehydrated through the alcohol series as before. These squashes were then hybridised to 3H-labelled pM56 as described for untreated squashes. DNase treatment of squashes. Another parallel series of squashes on slides was pretreated with DNase as follows: after fixation and dehydration, 50 ~1 of 100 pg/ml RNase-free DNase in 10 mM Tris, pH 7.4, was placed on each squash, covered with a glass coverstop, and incubated at 37°C for 30 min. Slides were washed twice for 15 min in 2~ SSC and then dehydrated through the alcohol series as before. These squashes were then hybridised to 3H-labelled pM56 as described for untreated squashes.
in
Non-lens
95
Tissues RESULTS
3H-Labelled pM56 was hybridised to squashes of tissue from 4- to 6-day lens, neural retina, retina pigment epithelium (not shown), adenohypophysis, epiphysis, otic vesicle, and heart. Parallel squashes were hybridised to 3H-labelled pBR322, the plasmid moiety of pM56. After a 6-week exposure, lens cells were all uniformly labelled by pM56 over cytoplasm and nuclei (data not shown). All non-lens tissues contained clusters of heavily labelled cells, with grains predominantly over the nuclei, surrounded by large areas of unlabelled cells with morphology and staining properties similar to the labelled cells (Fig. 1). Control squashes of all tissue types hybridised to 3H-labelled pBR322 contained no labelled cells (5-day lens only shown, Fig. 2D). Control squashes which were pretreated with RNase before hybridisation were negative (Fig. 2C), but DNase pretreatment did not prevent hybridisation (Fig. 2B). While labelled and unlabelled cells within a tissue were similar, they were morphologically distinct from cells in other tissues. Thus in each non-lens tissue examined, cells characteristic of that tissue were divided into two classes, a small class transcribing large amounts of &crystallin RNA, and a large class which did not transcribe the 6-crystallin genes, at least at a level detectable by in situ hybridisation. DISCUSSION
Since the plasmid vector pBR322 gave no hybridisation to squashes, and pM56 hybridises with high specificity to the two d-crystallin genes only, the heavy nuclear labelling of clusters of cells is due not to nonspecific binding, but to the presence of very large amounts of &crystallin RNA in the nuclei of these clusters. This is confirmed by the abolition of hybridisation by pretreatment with RNase. The absence of silver grains over cells with morphology similar to the labelled cells also indicates that labelling is not due to nonspecific sticking to some types of cells. We cannot precisely quantitate the number of transcripts, since different preparations retain accessible RNA to different extents, but with the fixation methods used here and the specific activity of the probe, several thousands of transcripts would be required to give the level of labelling observed (Hafen et ah, 1983; Angerer and Angerer, 1981). Since there are only 4-8 copies of the a-crystallin genes per diploid cell, and in any case only a few cells are labelled, we can rule out the possibility that the probe was hybridising to DNA. Furthermore the hybridisation is unaffected by pretreatment with DNase. The lower levels of labelling of cytoplasm might be due to some spreading from the nuclei during squashing, but data from Northern transfers (Bower et al, 1983b)
96
DEVELOPMENTAL BIOLOGY
VOLUME 112, 1985
FIG. 1. Embryonic tissues were dissected out, squashed, fixed, hybridised to ‘H-labelled pM56, dipped in Ilford K-2 nuclear emulsion diluted 1:l with HzO, dried, and exposed for 6 weeks at 4°C in a light-proof box. The original magnifications of the photographs are given. (A) 3.5- to 4-day neural retina, x400; (B) 5-day heart, X400; (C) 3.5- to 4-day epiphysis, X1000; (D) g-day adenohypophysis, X1000; (E) 3.5- to $-day otic vesicle, X1000; (F) g-day neural retina, X400.
indicate that while most of the RNA in non-lens tissues hybridising to pM56 (Scrystallin cDNA clone) was larger than the major d-crystallin mRNA, now known from sequence data to be 15’74nucleotides plus poly(A) (Yasuda et al., 1984; Nicker-son and Piatigorsky, 1984), a small percentage of the hybridising RNA in non-lens
tissues comigrated with the major lens Scrystallin mRNA. In addition, the size classes of RNA which hybridised to pM56 were different between nuclear and cytoplasmic RNA from non-lens tissues (Bower et ah, 1983b), suggesting that the cytoplasmic material detected was genuine and was not due to contamination
JEANNY ET AL.
6Gystallin
in Nm-lens
Tissues
FIG. 2. Embryonic 3.5-day lenses were treated and hybridised as for Fig. 1. (A) Hybridised to 3H-labelled pM56 as for Fig. 1, X1000. (B) Squash pretreated with DNase before hybridisation, X400. Hybridisation has not been prevented. (C) Squash pretreated with RNase before hybridisation, X400. (D) Squash hybridised to 3H-labelled pMB322, the plasmid moiety of pM56, X400.
by nuclear RNA. This implies that the light labelling detected over cytoplasm of cells with heavily labelled nuclei in Figs. lA, C, and D may be due to b-crystallin RNA being transported to the cytoplasm. Significant cytoplasmic label was not seen in other tissues (Figs. 1B and E) which are not known to have transdifferentiation potential. Low levels of crystallin antigenicity have been detected in all tissues of the eye (Maisel and Harmison, 1963; Clayton et ab, 1968; Bours and van Doorenmaalen, 1972) and at very low, or trace levels in several extraocular tissues including brain, liver, kidney, muscle, and skin (Clayton et al., 1968). More recently, immunohistology has shown cells with &crystallin antigenicity in adenohypophysis brain (Barabanov, 19’77, 1982) and epiphysis (Ueda et aZ., 1983; Watanabe et al., 1983). The detection of crystallin specificity does not, however, necessarily imply the presence of a completed polypeptide (Clayton et ah, 1968; Clayton, 1978), but these data suggest that some translation may be taking place on these templates.
Schibler et al. (1983) have found a similar pattern of expression in the a-amylase la gene, which is expressed in several tissues, including the brain, where RNA transcripts are confined to the nucleus, although it is as abundant as in liver, where the RNA is translated. They have not yet established whether all brain cells transcribe the gene, or only a subset, as we find here for the &crystallin genes. Mouse antitrypsin and antichymotrypsin genes, expressed at a high level in liver, are expressed at a low level in other tissues (Shaw et ah, 1983). Again, it is not known if expression is at a high level in a few cells, as we find for &crystallin genes, or at a low level throughout. We do not know whether the pattern of &crystallin gene transcription is the same in expressing non-lens cells as lens. It may be that of the two crystallin genes, a different one is transcribed in lens compared with other tissues. However, while &crystallin protein detected in NR is indistinguishable from that in the lens, the 6crystallin protein detected in the adenohypophysis differs electrophoretically from lens &crystallin (de Pom-
98
DEVELOPMENTAL BIOLOGY
erai and Clayton, 1978; Barabanov, 1977). If different genes are transcribed in different tissues, this would parallel the case of the a2u-globulin genes (Laperche et ah, 1983) and the MUP genes (Shaw et aZ.,1983). Another possibility is that the promoter used is different, as with the a-amylase la gene (Schibler et al., 1983), the ovalbumin gene (Malek et aZ.,1981), and the lysozyme gene (Grez et ah, 1981). Our present data show that the predominantly highmolecular-weight d-crystallin RNA transcripts previously detected in tissues unrelated to the eye (Bower et al., 1983a) and in eye-associated tissues with transdifferentiation capacity are due to high-level transcription in a few cells rather than low level in all. Heterogeneity with respect to methylation of the Scrystallin genes has also been observed in g-day NR (Cooper et ah, 1983). In NR a specifically located subpopulation appears to be involved in d-crystallin transcription (Bower et ab, 1983b), but in other tissues it is not yet known whether this is the case; no morphological or staining difference has been found between labelled and unlabelled cells, suggesting that they may all be cells of the same subtype. The clustering of labelled cells in the tissues examined suggests either that cells in a cluster are clonally related or that they are in a special location in which an environmental factor induces Scrystallin gene expression. We are currently investigating these possibilities. Higher levels of RNA detected in NR compared with heart are probably due to a higher proportion of cells expressing the genes in NR, although we do not yet have precise figures for the same embryonic stage in both tissues. On serial sections of the central part of the 3.5 day eye, 15% of NR cells were labelled (unpublished data), and in squashes of the whole &day heart, 0.1% of cells were labelled. The number of labelled cells in a cluster was higher for neural retina than for heart. However, we do not have figures for the entire, total 3.5day NR, and it is possible that there is some difference between different parts of the retina, since the pattern of labelling is not random, but localised to the vitreal border. In addition, there may be some change between 3.5 and 5 days in percentage of cells labelled. There appear to be cells, sometimes within a single tissue, exhibiting several different transcriptional states of the Scrystallin genes: those in which it would appear that these crystallin sequences are not transcribed at detectable levels, those in which they are transcribed at high levels and rapidly processed, as in embryo lens, and those in which there is appreciable transcription as in the non-lenticular tissues examined here, some of which appear also to show some degree of processing. The level of processing may be higher in retina than in otic vesicle, and we present elsewhere evidence that retina cells have
VOLUME 112, 1985
latent processing capacity (Clayton et aL, in press; Bower et al., submitted for publication). We are now looking for expression of other crystallin genes in these tissues, and investigating the fate of crystallin gene-expressing clusters in development. This work was supported by the Medical Research Council. J-C. Jeanny was supported by a Royal Society Visiting Research Fellowship. We are grateful to the Poultry Research Centre, Roslin, Midlothian, for fertile eggs.
REFERENCES AGATA, K., YASUDA, K., and OKADA, T. S. (1983). Gene coding for a lens-specific protein, 8-crystallin, is transcribed in non-lens tissues of chicken embryos. Dev. Biol. 100,222-226. ANGERER, L. M., and ANGERER, R. C. (1981). Detection of poly A + RNA in sea urchin eggs and embryos by quantitative in situ hyhridisation. Nucleic Acids Res. 9, 2819-2840. BARABANOV, V. M. (1977). Detection of &crystallins in the adenohypophysis of the chick embryo (Russian). Dokl. Akcd. Nauk. SSSR 234,195-198. BARABANOV, V. M. (1982). Extra lenticular localisation of b-crystallin in Japanese quail (Russian). Dokl. Akad. Nauk. SSSR. 262,22-24. BHAT, S. P., and PIATIGORSKY, J. (1979). Molecular cloning and partial characterisation of 6-crystallin cDNA sequence in a bacterial plasmid. Proc. NatZ. Acad. Sci. USA 76,3299-3303. BOURS, J., and VAN DOORENMAALEN, W. J. (1972). The presence of lens antigens in intraocular tissues of the chick eye. Exp. Eye Res. 13, 236-247. BOWER, D. J., ERRINGTON, L. H., WAINWRIGHT, N. R., SIME, C., MORRIS, S., and CLAYTON, R. M. (1982). Cytoplasmic RNA sequences complementary to cloned chick &crystallin cDNA show size heterogeneity. Biochem. J 201,339-344. BOWER, D. J., ERRINGTON, L. H., POLLOCK, B. J., MORRIS, S., and CLAYTON, R. M. (1983a). The pattern of expression of chick &crystallin genes in lens differentiation and in transdifferentiating cultured tissues. EMBO J. 2, 333-338. BOWER, D. J., ERRINGTON, L. H., COOPER,D. N., MORRIS, S., and CLAYTON, R. M. (1983b). Chicken lens &crystallin gene expression and methylation in several non-lens tissues. Nucleic Acids Res. 11,2513-2527. BOWER, D. J., JEANNY, J-C., ERRINGTON, L. H., MORRIS, S., and CLAYTON, R. M. (1985). A quantitative comparison of &crystallin gene expression during development of chick lens and non-lens tissues. Submitted for publication. CLAYTON, R. M. (1978). Divergence and convergence in lens cell differentiation: Regulation of the formation and specific content of the lens fibre cell. In “Stem Cells and Tissue Homeostasis,” (B. I. Lord, C. S. Potten, and R. J. Cole, eds), C.U.P., pp 115-138. CLAYTON, R. M. (1982). The molecular basis for competence, determination and transdifferentiation: A hypothesis. A&J. Exp. Biol. Med. 158,23-38. CLAYTON, R. M., CAMPBELL, I. C., and TRUMAN, D. E. S. (1968). A reexamination of the organ specificity of lens antigens. Ezp. Eve Res. 7, 11-29. CLAYTON, R. M., JEANNY, J-C., BOWER, D. J., and ERRINGTON, L. H. (1985). Extra lenticular crystallins and their role during transdifferentiation to lens. Curr. Top Dev. Bid, in press. COOPER,D. N., ERRINGTON, L. H., and CLAYTON, R. M. (1983). Variation in the DNA methylation pattern of expressed and nonexpressed genes in chicken. DNA 2,129-138.
JEANNY ET AL.
6-Crystallin
GREZ, M., LAND, H., GIESECKE, K., SCHUTZ, G., JUNG, A., and SIPPEL, A. E. (1981). Multiple mRNAs are generated from the chicken lysozyme gene. Cell 25,743-752. HAFEN, E., LEVINE, M., GARBER, R. L., and GEHRING, W. J. (1983) An improved in situ hybridisation method for the detection of cellular RNAs in Drosophila tissue sections and its applications for localising transcripts of the homeotic Antennapedia gene complex. EMBO J. 2, 617-623. JACKSON, J. F., CLAYTON, R. M., WILLIAMSON, R., THOMSON, I., TRUMAN, D. E. S., and DE POMERAI, D. I. (1978). Sequence complexity and tissue distribution of chick lens crystallin mRNAs. Dev. Biol 65, 383-395. LAPERCHE, Y., LYNCH, K. R., DOLAN, K. P., and FEIGELSON, P. (1983). Tissue-specific control of 012~globulin gene expression: Constitutive synthesis in the sub-maxillary gland. Cell 32,453. MAISEL, H., and HARMISON, C. (1963). An immunoembryological study of the chick iris. J. EmbyoL Exp. MorphoL 11,483-491. MAITLAND, N. J., KINROSS, J. H., BUSU~~IL, A., LUDGATE, S. M., SMART, G. E., and JONES, K. W. (1981). Detection of DNA tumour virusspecific RNA sequences in abnormal human cervical biopsies by in situ hybridisation. J. Gen. ViroL 55, 123-137. MALEK, L. T., ESCHENFELDT, W. H., MUNNS, T. W., and RHOADS, R. E. (1981). Heterogeneity of the 5’terminus of lens ovalbumin mesenger ribonucleic acid. Nucleic Acids Res. 9, 1657-1673. MANIATIS, T., JEFFREY, A., and KLEID, D. G. (1975). Nucleotide sequence of the rightward operator of phage lambda. Proc. NatL Acad Sci. USA 72.1184-1188. NICKERSON, J. M., and PIATIGORSKY, J. (1984). Sequence of a complete chicken 6-crystallin cDNA. Proc. NatL Acad. Sci. USA 81,2011-2015. OKADA, T. S. (1983). Cellular and molecular aspects of transdifferentiation of ocular tissues in vitro. Cell &%r. 12, 117-184. PIATIGORSKY, J., BEEBE, D., ZELENKA, P., MILSTONE, L. M., and SHINOHARA, T. (1976). Regulation of &crystallin gene expression during development of the embryonic chick lens. INSERM 60,85-112. DE POMERAI, D. I., PRITCHARD, D. J., and CLAYTON, R. M. (1977). Biochemical and immunological studies of lentoid formation in cultures
in Nor-lens
Tissues
99
of embryonic chick neural retina and day-old chick lens epithelium. Dev. BioL 60,416-427. DE POMERAI, D. I., and CLAYTON, R. M. (1978). Influence of embryonic stage on the transdifferentiation of chick neural retina cells in culture. J. EmbryoL Exp. MorphoL 47, 178-193. SANDRITTER, W., PILNY, V., NOVAKOVA, V., and KIEFER, G. (1966). Zur Problematik der Gewebspraparation fur cytophotometrische Messungen. Histochemie 7, l-7. SCHIBLER, U., HAGENBUCHLE, O., WELLAUER, P. K., and PIITET, A. C. (1983). Two promoters of different strength control the transcription of the mouse alpha-amylase gene Amy-la in the parotid gland and the liver. Cell 33, 501-508. SHAW, P. H., HELD, W. A., and HASTIE, N. D. (1983). The gene family for major urinary proteins: Expression in several secretory tissues of the mouse. Cell 32,755-761. THOMSON, I., YASUDA, K., DE POMERAI, D. I., CLAYTON, R. M., and OKADA, T. S. (1981). The accumulation of lens-specific protein and mRNA in cultures of neural retina from 3.5 day chick embryos. Exp. Cell Res. 135, 445-449. UEDA, G., HIRO, F., TAKEICHI, M., and OKADA, T. S. (1983). Detection of 6-crystallin in adenohypophysis in chick embryos using monoclonal antibodies. Dev. Growth Difler. 25,421. WATANABE, K. (1984). In “Proceedings, 12th NIBB Conf. Molecular Mechanisms of Cell Specialisation in Development.” WATANABE, K., TAMAMAKI, N., and NOJYO, Y. (1983). Probable presence of oculopotent cells in embryonic pineal of quail. Dev. Growth Difer. 25, 421. YASUDA, K., KONDOH, H., OKAZAKI, T. S., OKADA, T. S., and SHIMURA, Y. (1982). Organization and structure of (Y- and &crystallin genes and their expression in cells by microinjection. In “Proceedings, Vth Int. Cong. Eye Res., Eindhoven,” p. 38. YASUDA, K., NAKAJIMA, N., ISOBE, T., OKADA, T. S., and SHIMURA, Y. (1984). The nucleotide sequence of a complete chicken &crystallin cDNA. EMBO J. 3,1397-1402. ZWAAN, J., and IKEDA, A. (1968). Macromolecular events during differentiation of the chick lens. Exp. Eye Res. 7,301-311.