THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 267, No. 21. Issue of July 25, pp. 15105-15110,1992 Printed in U. S. A.

Tissue- and Development-specific Expression of the Human Phenylalanine Hydroxylase/Chloramphenicol Acetyltransferase Fusion Gene in Transgenic Mice*

(Received for publication, January 15, 1992)

Yibin WangS, Janet L. DeMayoS, Tina M. HahnS, Milton J. Finegoldg, David S. Koneckill, Uta Lichter-Koneckill, and Savio L. C. WooS)I From the $Howard Hughes Medical Institute, Department of Cell Biology, §Department of Pathology, Baylor College of Medicine, Houston, Texas 77030 and the 7lUniuersitats-Kinderklinik, im Neuenheimer Feld 150, W-6900 Heidelberg, Germany

Human phenylalanine hydroxylase (PAH) catalyzes the conversion of L-phenylalanine to L-tyrosine. Defi- ciency of this enzyme results in phenylketonuria, a common genetic disorder of amino acid metabolism that causes severe mental retardation. In primates, PAH is expressed specifically in the liver, while in rodents PAH activity is also present in kidney, al- though at a much lower level. A 9-kilobase genomic DNA fragment at the 5’ end of the hPAH gene (hPAH) was fused to the bacterial chloramphenicol acetyl- transferase (CAT) gene. The hPAH/CAT minigene was used to generate multiple transgenic mouse lines. In all expressing lines, CAT activity was detected predomi- nantly in the liver and at much lower levels in the kidney. By immunohistochemical staining, CAT expression was localized to hepatocytes and renal epi- thelial cells, both of which also express the endogenous mouse PAH enzyme. Furthermore, both the transgene and the endogenous mouse PAH were activated at about the same stage of embryonic development in the mouse liver. These results suggest that the 9-kilobase DNA fragment flanking the 5’ end of the human PAH gene contains all the necessary cis-acting elements to direct tissue- and developmental-specific expression in vivo.

Human hepatic phenylalanine hydroxylase (PAH,’ phen- ylalanine 4-monooxygenase, EC 1.14.16.1) is a mixed-function monooxygenase that catalyzes the hydroxylation of L-phen- ylalanine to L-tyrosine (1). Deficiency of this enzyme results in phenylketonuria (PKU), an autosomal recessive disorder with an average incidence of 1 case/10,000 Caucasian live births (2). Previously, we reported the isolation of the cDNA clone phPAH247 (4) which was later shown to contain all the

* This work was supported in part by a grant from the National Institutes of Health HD-17711 (to S. L. C. W.) and in part by Deutsche Forschungsgemeinschaft Grant Li.375/2-1/2-2 (to U. L-K.) and Fritz Thyssen Foundation Grants 1990/51 (to Prof. H. J. Bremer, U. L-K., and D. S. K.) and 8.11/91 (to D. S. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduer- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

11 Investigator of the Howard Hughes Medical Institute. To whom correspondence should be addressed Howard Hughes Medical Insti- tute, Baylor College of Medicine, Houston, TX 77030. Tel.: 713-798- 6080; Fax: 713-790-1211.

The abbreviations used are: PAH, phenylalanine hydroxylase; CAT, chloramphenicol acetyltransferase; PKU, phenylketonuria; PCR, polymerase chain reaction; kb, kilobase.

genetic information necessary to code for the functional PAH enzyme (5). Using this cDNA, the human PAH/PKU locus was mapped to chromosome 12 (3). Subsequently, the genomic structure of the human PAH gene was also characterized (6). The cloning of the human PAH cDNA and characterization of the genomic structure has led to the discovery of numerous mutations in the PAH gene that cause PKU (for latest review, see Ref. 7). However, since the essential regulatory elements for the PAH gene expression have not been defined, potential mutations affecting the transcription of the PAH gene have not yet been identified.

The mature human PAH messenger RNA is approximately 2.4 kb in length and is encoded by 13 exons. The sequence of the 5’-flanking region shows many features characteristic of some house keeping genes, e.g. no obvious TATA box, multiple GC-rich domains, and multiple CAP sites.’ However, expres- sion of PAH follows a specific pattern in terms of tissue distribution and developmental ontogenesis. In humans and other primates, PAH was reported to be expressed specifically in liver (9), although one recent report observes a very low level of PAH mRNA in leukocytes (8). In rodents, PAH activity was also detected in kidney at about 15% of the activity in liver (10 , l l ) . During embryonic development, PAH activity was first detected in rat liver at very late stages of development (day 21 of gestation ( l l ) , (15)) while it was detected in human fetal liver tissue, as early as the first trimester (12, 13). Both in terms of tissue-specific distribution and developmental onset, the detection of PAH enzyme activ- ity coincides closely with the presence of PAH mRNA (14, 15). These observations suggest that the tissue-specific and developmental stage-specific expression of the PAH gene may be regulated at the transcriptional level.

In liver-specific genes, such as albumin and a-1-antitrypsin, multiple cis-acting regulatory elements in the 5’-flanking region have been shown to be responsible for tissue specificity (16-18). Numerous transcriptional factors interacting with these specific DNA elements, including HNF-1 (19), HNF-3a ( Z O ) , and C/EBP(Zl), have been characterized, and their genes have been cloned. No cis-elements responsible for tissue or developmental stage specificity have, yet, been identified in the PAH gene. In this report, we employed transgenic mice as an in vivo model to define the regulatory functions of a 9- kb 5’-flanking DNA fragment of the human PAH gene. The results suggest that this DNA fragment contains cis-elements required for tissue-specific and developmental stage-specific expression in vivo.

* D. Konecki, manuscript in preparation.

15105

15106 Human Phenylalanine Hydroxylase Gene Expression in Transgenic Mice

MATERIALS AND METHODS

Cells and Culture Conditions-HepG2, a human hepatoma cell line (obtained from America Type Culture Collections, ATCC HB 8065) was cultured as described by Kelly et al. (30) in 25% Waymouth medium and 75% modified Eagle's medium, supplemented with 10% fetal bovine serum. Folse cells, a primary human fibroblast (kindly provided by Dr. F. D. Ledley, Baylor College of Medicine), were grown in modified Eagle's medium supplemented with 20% fetal bovine serum, and 2.5 X amino acid solutions (GIBCO 320-1130AG), and 1 X vitamin solution (GIBCO 320-1120AG). Both cell lines were in- cubated at 37 "C in the presence of 5% CO,. Upon reaching 290% confluence, they were harvested for RNA preparation.

Construction of the hPAH/CAT Fusion Gene-An approximately 9-kb human PAH genomic fragment containing the 5"flanking region and part of the first exon was isolated from a cosmid clone, cPAH15 (6), by EcoRI and SmaI digestion. The fragment was then inserted into the Hind111 site of the pSVoCAT vector (22) after being blunt- ended by DNA polymerase (Klenow fragment). The 11.2-kb hPAH/ CAT fusion gene containing the 9-kb human PAH fragment, the chloramphenicol acetyltransferase-coding sequences, a 3"splicing signal, and a poly(A) signal from SV40 was isolated by PstI and Hue11 digestion (Fig. 1) and was subsequently used to create transgenic mice.

Generation and Analysis of Transgenic Mice-The DNA fragment was isolated from a 0.8% low melting agarose gel and purified by Geneclean (BiolOl, La Jolla, CA). The microinjection of the DNA fragment into the male pronuclei of mouse embryos (ICR X B6D2F,) was performed according to the procedure of Brinster et al. (23) as described by Sifers et al. (24). Transgenic mice were identified by Southern blot analysis of tail DNAs isolated from mice weaned at 21 days of age, as previously described (24). In brief, 5 pg of tail DNA was digested to completion by BamHI restriction enzyme (Promega), fractionated in 0.8% agarose gels, and transferred to nitrocellulose filters. The filters were hybridized with a probe prepared from the PstIIHaeII fragment by random primed labeling and washed to a stringency of 0.2 X SSC, 0.5% sodium dodecyl sulfate at 68 "C.

CATAssay-The various mouse tissues were homogenized in 0.25 M Tris-HC1, pH 7.5, buffer and denatured as previously described (18). 1-100 pg of denatured protein was incubated with 80 pg of acetyl-coA (Sigma) and 0.5 pCi of [14C]chloramphenicol (Du Pont- New England Nuclear) at 37 "C for 20 min. The assays were per- formed within the linear range of the reaction, and the conversion percentage was calculated from the acetylated products uersus total input of the [14C]chloramphenic~l substrate in the reaction.

Immunohistochemical Staining-Tissues from the adult transgenic mice were fixed in ice-cold absolute ethanol and prepared for immu- nohistochemical staining as described (32). The primary antibody used on the sections were goat anti-rat PAH antibody (14), and rabbit anti-CAT antibody (5 Prime-3 Prime Inc., Paoli, PA). Standard immunoperoxidase procedures were then performed as described (32).

Embryonic Studies-Fetuses at different gestational ages were isolated from pregnant ICR female mice mated with a hPAH/CAT10 transgenic male. Minimal amounts of fetal tissues were used to extract DNA for determination of genotype by polymerase chain reaction (PCR). The PCR was performed as described (25), using approxi- mately 2 pg of fetal genomic DNA and 0.6 pg of a pair of 20-mer oligonucleotide primers complementary to the human PAH gene, amplifying a fragment from -166 to +174 relative to the first initia- tion site.3 Liver homogenates from the transgene-positive fetuses were then assayed for CAT activity, as described earlier, and for PAH activity by measuring the conversion of ~-['~C]phenylalanine to L- [14C]tyrosine as described elsewhere (5).

Primer Extension-Total RNA was isolated from mouse liver tissue by the method of Chirgwin et al. (26) and from human HepG2 cells and primary fibroblasts by the hot-phenol method (27). The poly(A) RNA was prepared from total RNA using oligo(dT)-cellulose according to Gilham (28). The details of the primer extension have been described (18). The primer used on the mouse liver RNA was a synthetic 30-mer oligonucleotide complementary to the sequences coding for amino acids 6-15 of the CAT protein (29). The primer used on HepG2 and fibroblast RNA was a 21-mer complementary to the first 7-amino-acid coding region of the human PAH protein (4).

Y. Wang, unpublished data.

RESULTS

Generation of the PAHICAT Transgenic Mice-The entire human PAH gene has been cloned into four overlapping cosmid clones (6), and the clone cPAH15 contains the most 5' portion of the gene. As the first approach to localize the functional cis-acting regulatory elements, an approximately 9-kb EcoR I to SmaI fragment, including part of the exon 1 and the 5"flanking region, was excised from cPAH15. This fragment was then fused with a CAT gene and SV40 poly(A) signal sequences to construct the hPAH/CAT fusion gene (Fig. 1). Subsequently, the chimeric hPAH/CAT gene was introduced into the mouse genome to create transgenic mice. The Fo mice bearing the hPAH/CAT gene were identified by Southern blot screening of tail DNA (data not shown). From the transgene positive Fo mice, four transgenic lines were established. They were designated as hPAH/CATl, hPAH/ CAT5, hPAH/CAT8, and hPAH/CATlO, each carrying ap- proximately 1, 5, 8, and 10 copies of the transgene, respec- tively.

Tissue Distribution of CAT Activity in Transgenic Mice- Protein from tissues of the four transgenic mouse lines was extracted and assayed for CAT activity to determine the tissue distribution. As illustrated in Fig. 2, the highest CAT activity was detected in liver from all four transgenic lines. However, there was no apparent correlation between the CAT activity in the liver and the copy number of the transgene among different lines. Significant CAT activity was also detected in kidney from all the lines, ranging from 10 to 35% relative to the activity in the liver. The consistent expression of CAT in the liver and kidney coincided with the endogenous mouse PAH expression pattern as previously reported (10, 11). No significant CAT activity was consistently detected from other tissues.

Cell Type Specificity of CAT Expression in Transgenic Mice-Liver, kidney, and 11 other tissue samples (as listed in Fig. 2) were then analyzed by histochemical staining to local- ize cell type specificity of the CAT gene expression in the transgenic mice. Sections of liver and kidney from the hPAH/ CAT10 transgenic line exhibited positive staining as shown in Fig. 3. The CAT protein was detected in almost all hepa- tocytes, although at very different levels, as well as, in about 1-5% of the proximal tubule cells of the kidney (Fig. 3, D and F ) . No positive staining was detected in other tissue samples (data not shown). The liver sections from hPAH/CAT5 and hPAH/CAT8 lines showed a similar staining pattern, while

-9ooo Human PAH vector

I !

1 EroR I I I

I I - 5.6 hilobaleb I

Sma 1

1 I BamH I

I, Barn111 '

13.X kilobases -1 l lae I 1 1 P*l I

FIG. 1. A schematic structure of the hPAH/CAT fusion gene. An approximately 9-kb EcoRI to SmaI fragment, covering the 5'- flanking region and part of the 5'-untranslated region in exon 1, was fused with the CAT gene and SV40 poly(A) signal sequences. The 11.2-kb fusion gene fragment was removed from the vector by PstI and Hue11 digestion before microinjection into the male pronuclei of fertilized mouse eggs to create transgenic mice. 46 FO mice were screened by Southern blotting of BamHI-digested tail DNA. Trans- gene positive mice were identified by the presence of a 5.6-kb band. From the transgene positive Fa mice, four transgenic lines were established and analyzed in this report. Each has approximately one, five, eight and 10 copies of the transgene and was designated as hPAH/CATl, hPAH/CAT5, hPAH/CAT8, and hPAH/CAT10, re- spectively.

Human Phenylalanine Hydroxylase Gene Expression in Transgenic Mice 15107

FIG. 2. Tissue-specific expression of CAT in transgenic mice. Protein homogenates in 0.25 M Tris-HC1, pH 7.5, from different tissues were incubated in a 100-pl final volume with 0.5 pCi of ['4C]chloramphenicol, and 4 mM acetyl- CoA at 37 "C for 20 min. The reactions were performed in the linear range, using 1 pg of protein from the hPAH/CAT10

CAT5 (stripe), and hPAH/CAT8 (solid), 5 pg of protein from hPAH/

from hPAH/CATl line (blank). The (shadow) lines and 100 pg of protein

CAT activities (mean value of two or three sets of samples from each line) are

being normalized for the different shown as percentage conversion after

amounts of protein used in the assay (% conversion/pg protein). The error bars represent standard deviations.

40 - I

0.2 -.

.T

0 1

0 0

liver sections from hPAH/CATl line showed no positive staining due to the low level of CAT expression in that line (data not shown). Comparing these results with the endoge- nous mouse PAH expression detected by immunochemical staining using an anti-rat PAH antibody (14) (Fig. 3, C and E) , the expression of the transgene coincided with that of the endogenous gene in the same cell types both in liver and kidney. Quantitatively, however, the endogenous PAH gene showed a much more homogeneous expression level among all hepatocytes.

CAT Expression in Transgenic Mouse Liver Was Directed by the Functional Promoter of the Human PAH Gene-The tissue-specific and cell type-specific expression of CAT in multiple transgenic lines strongly suggested that the trans- gene was under the regulation of the human PAH gene fragment. To provide further evidence for the function of the human PAH 5'-flanking sequence in the transgenic mice, the transcriptional initiation site of the fusion gene was mapped by primer extension. Using both primer extension and S1 nuclease protection methods, our laboratory demonstrated that the human PAH gene may have multiple CAP sites, with the most 5' site located 154 base pairs upstream of the translation initiation codon.2 Total and poly(A)+ RNA iso- lated from transgenic mouse liver of the hPAH/CAT10 line were analyzed by primer extension using an oligonucleotide complementary to the CAT protein coding region. The results, as shown in Fig. 4, demonstrated that the most upstream CAP site (Cl ) and other major downstream CAP sites (C4, C5, and C6) detected in the transgenic mouse liver were colocalized with the CAP sites utilized by the PAH gene in

human hepatoma (hepG2) cells. Therefore, the CAT gene expression in the liver was indeed driven by the functional human PAH promoter.

CAT Expression in Transgenic Mouse Liver during Embry- onic Development-During embryonic development, the ele- vation of CAT activity in the liver was first detected at approximately day 17-18 of gestation (Fig. 5). The expression level was induced dramatically within 1 week after birth. The endogenous mouse PAH activity was also measured from the same set of samples. There was some activity detected a t 15- 17 days, but significant induction occurred only after 18 days of gestation. A surge of PAH activity after birth was also observed. The parallel developmental pattern of the transgene and the endogenous mouse PAH gene indicated that the 9-kb human PAH fragment contained regulatory elements respon- sible for the developmental regulation of the PAH gene.

DISCUSSION

Multiple lines of transgenic mice carrying a fusion gene with a 9-kb human PAH 5'-flanking DNA fragment driving the reporter gene CAT were created to investigate the regu- latory functions of the 5'-flanking region of the human PAH gene. Three lines of evidence suggested that the 9-kb fragment functioned in vivo to direct tissue-specific and development- specific expression: 1) consistent expression of CAT in the liver and kidney among all transgenic lines; 2) close correla- tion between the transgene and the endogenous mouse PAH gene in developmental specific activation and cell type-spe- cific distribution; 3) colocalization of major transcriptional initiation sites utilized by the transgene in the mouse liver

15108 Human Phenylalanine Hydroxylase Gene Expression in Transgenic Mice

”

FIG. 3. Cell type-specifh expression of the CAT gene uersm mouse PAH gene. Immunoperoxidase staining for mouse PAH (C and E ) and CAT (B, D, and F ) was accomplished using anti-rat PAH and anti-CAT antibodies, respectively, as described under “Materials and Methods.” The tissue sections used for staining were from liver (A, X200), (C, X400), (D, ~ 6 4 0 ) , and kidney (E, X400), (F, X600) of the hPAH/CAT10 line. A is the technical control without specific primary antibodies. B, X200 is the negative control on a liver section from non-transgenic mouse liver. There was no positive staining for CAT and PAH from sections of other tissues (data not shown).

and by the PAH gene in human hepatoma cells. It has been established that multiple cis-acting promoter

and enhancer sequences are required to direct specific expres- sion of genes (38-46). These cis-elements can be recognized by cellular transcription factors, which bind to the corre- sponding DNA sequences to initiate the transcription and/or to regulate the transcription rates (46). The tissue-specific expression of many hepatic genes has been shown to result from the combined effects of multiple protein factors inter- acting with multiple cis-elements. These include tissue-spe-

cific or tissue-enriched transcription factors, such as HNF-1 (19), CBBP (21), and HNF-3a (20), and ubiquitous factors, such as AP-1 (34), SP-1 (33), and NF-1 (35). The tissue- specific regulation observed in vivo is suggested to be achieved by balancing the different transcription factors among differ- ent tissues (46,47). In fact, the difference between the patchy expression of CAT and the homogenous expression of the endogenous mouse PAH (compare Fig. 3, C and D) suggests that this human PAH gene fragment may still lack some cis- elements required for precise regulation of the PAH gene.

Human Phenylalanine Hydroxylase Gene Expression in Transgenic Mice

5 1

20 1-

1.30-

180-

1.30-

147-

\ $48- - 211"

1 . 3 0 "

1 w " .

L 1 IO"

"

1 l D -

00-

"

- 1

-. 1"

I 1 I I -100 .I .I40 ' 2 2 0

FIG. 4. Primer extension to determine the transcription ini- tiation sites of the hPAH/CAT gene in the transgenic mouse liver. A, 25 pg of total and 10 pg of poly(A)+ RNA (lanes 2 and 3 ) isolated from the transgenic liver of hPAH/CAT10 line were hybrid- ized to a 30-mer oligonucleotide complementary to the CAT-coding sequences (CAT primer), and extended with AMV reverse transcrip- tase. Primer extension was also performed on 25 pg of total and 10 pg of poly(A)+ RNA (lanes 4 and 5) isolated from human hepatoma HepG2 cells using a primer complementary to the first seven human PAH coding sequences (PAH primer). The multiple CAP sites utilized in the transgenic mouse liver were aligned to the same CAP sites as observed in HepG2 cells. Lanes I and 6 were primer extension on 10 pg of poly(A)+ RNA from non-transgenic mouse liver and human fibroblasts using the CAT and PAH primers, respectively. B, sche- matic structure of the CAP sites of human PAH gene in HepG2 cells and the hPAH/CAT fusion gene in transgenic mouse liver.

The persistence of such CAT expression patterns among multiple transgenic mouse lines argued against the hypothesis that the non-homogenous expression could be caused by in- sertional effects. Identification of the regulatory cis-elements present within the 5"flanking region and the corresponding trans-acting factors for the PAH gene expression is currently the subject of further investigation.

This study presented the first evidence that PAH expres- sion in mice was restricted to a specific cell type in the kidney. Even though renal PAH activity in rodents has been investi- gated for decades, its functional significance is still unclear. The localization of the PAH protein to the proximal tubule epithelial cells may provide additional insight into the role of renal PAH activity in gluconeogenesis and phenylalanine homeostasis in rodents, as proposed by Rao and Kaufman (31). It is also interesting to note that only a small percentage of the tubular cells expressed CAT and mouse PAH a t de- tectable levels, indicating that kidney tubular cells may have different functional properties. The exact mechanism for this phenomena needs to be studied further.

The PAH expression pattern has distinct species specificity.

f

1 1 1 3 1 6 1 7 I 8 1 9

Days of Gestation

I 15109

l a r r k

Aher Birth

FIG. 5. The expression of CAT versus mouse PAH during embryonic development in the liver. Fetuses carrying the trans- gene (as detected by PCR) from the hPAH/CATlO line were removed from the uterus a t different gestational stages. Liver protein extract was prepared as described in the legend to Fig. 2. The PAH activity (50 pg of protein, shadowed bars) was assayed by measuring the conversion of [14C]phenylalanine to [14C]tyrosine in the presence of tetrahydrobiopterin. The CAT activity (3 pg of protein, blank bars) was assayed as previously described. The conversion rates were cal- culated from the assay results performed on two sets of fetal liver samples, and each had two samples a t one time point. The error bars represent the standard deviation of the data.

In tissue distribution, humans and other primates do not express PAH in the kidney while rodents do have significant renal PAH activity (9, 14). During embryonic development, human PAH activity can be detected at early gestational stages (9,12), while the onset of the mouse PAH gene is much later, at about 1 or 2 days before birth (11, 15). The observed species specificity could result from either different cis-ele- ments of the promoter or different cellular transcription factor systems between one species and the other. In transgenic mice, the human PAH/CAT transgene closely followed the endogenous mouse PAH gene expression pattern rather than retaining the pattern of expression as in humans. Since the same DNA fragment might be responsible for the different expression patterns in humans and the transgenic mice, the species specificity of the PAH expression pattern is most likely due to the presence of different transcription factors in the kidney and the different development programming of the liver in humans and mice. Further characterization of the regulatory elements for the PAH gene may reveal the molec- ular basis for the different expression patterns among differ- ent species.

Finally, characterization of the regulatory elements for the PAH gene expression will provide a molecular basis for the identification of potential promoter mutations among mutant PAH alleles. Recently, naturally occurring mutations have been identified in the promoter region of the human Factor IX gene (48) and the retinoblastoma gene (49), causing re- duced expression of the corresponding genes. Although PAH promoter mutations have not, as yet, been described, there are PKU patients with no mutations in the coding region of their PAH genes." Therefore, we believe that the results reported here will not only provide basic guidance to further characterization of the regulatory elements, but also direct our pursuit of novel PKU mutations within the promoter region of the human PAH gene.

Acknowledgments-We thank Azita Reger, Xiao-Hong Chen, and Angela M. Major for their excellent technical assistance. We also thank Dr. Randy C. Eisensmith and colleagues in the laboratory of S. L. C. Woo for their critical discussions and constructive advice during the preparation of the manuscript.

15110 Human Phenylalanine Hydroxylase Gene Expression in Transgenic Mice REFERENCES

1. Kaufman, S. (1976) in Aduances in Neurochemistry (Agranoff, B. W. &

2. Scriver, C. R. & Clow, L. L. (1980) Annu. Reu. Genet. 1 4 , 179-202 Aprison, M. H., eds) Vol. 2., pp. 1-132, Plenum Press, New York

3. Lidsky, A. S., Law, M. L., Morse, H. G., Kao, F. T. &Woo, S. L. C. (1985)

4. Kwok, S. C. M., Ledley,, F. D., DiLella, A. G., Robson, K. J. H. & Woo, S.

5. Ledley, F. D., Grenett, H. E., DiLella, A. G., Kwok, S. C. M. & WOO, S. L.

6. DiLella, A. G., Kwok, S. C. M., Ledley, F. D., Marvit, J. & Woo, S. L. C.

7. Eisensmlth, R. C. and Woo, S. L. C. (1992) Hum. Mut. 1 , in press 8. Sarkar, G. & Sommer, S. (1989) Science 244 331-334 9. Hsieh, M. C. & Berry, H. K. (1979) J. Exp. Zool. 208 , 161-168

Proc. Natl. Acad. Sci. U. S. A. 8 2 , 6221-6225

L. C. (1985) Biochemt&y 24,556-561

C. (1985) Science 228 , 77-79

(1986). Biochemistry 2 5 , 743-749

10. Tourian, A., Goddard, J. & Puck, T. T. (1969) J. Cell. Physiol. 73,159-170 11. McGee, M. M., Greengard, 0. & Knox, W. E. (1972) Biochem. J . 127,669-

12. Greengard, 0. (1977) Pediatr. Res. 11 , 669-676 13. Raiha, N. C. R. (1973) Pediatr. Res. 7 , 1 14. Robson. K. J. H.. Chandra. T.. MacGillivrav. R. T. A. & Woo. S. L. C.

674

". ~ . . "" ~ . ~~ ~ ", ~ ~ ~

(1982) Proc. Natl. Acad. S c i U . S. A . 79,4761-4705

Dahl, H. M. (1988) Differentiation 3 8 , 42-48 15. Yeoh, G . C. T., Edkins, E., Mackenzie, K., Fuller, S., Mercer, J. F. B. &

16. Maire. P.. Wuarin. J. & Schihler. U. (1989) Scrence 244,343-346 17. Simone, V. D., Ciliherto, G., Hardon, E., Paonessa, G., Palla, F., Lundberg,

18. Shen. R. F.. Li. Y.. Sifers. R. N.. Wana. H.. Hardick. C., Tsai, S. Y. &Woo, L., & Cortese, R. (1987) EMBO J. 6,2759-2766

~ S. L. C. (1987) hucleicAcids kes. 13,83994415' 19. Frain, M., Swart, G., Monaci, P., Nicosia, A., Stampfli, S., Frank, R. &

20. Lai. E.. Prezioso. V. R.. Tao. W.. Chen, W. S. & Darnell, J. E., Jr. (1991) Cortese, R. (1989) Cell 5 9 , 145-157

Genes & Deu. 6,416-427

S. L. (1988) Genes & Deu. 2 , 786-800

. .

21. Landschulz, W. H., Johnson, P. E., Adash, E. Y., Graves, B. J. & McKnight,

22. Gorman. C. M., Moffat, L. F. & Howard, B. H. (1982) Mol. Cell. Biol. 2 , 1044

23. Brinster, R. L., Ritchie, K. A,, Hammer, R. E., O'Brien, R. L., Arp, B. &

24.

26. 25.

27. 28. 29. 30. 31. 32.

33. 34.

35. 36. 37. 38.

39.

40.

41. 42. 43. 44.

45.

46.

47.

48. 49.

Sifers, R. N., Carlson, J. A., Clift,,S. M., DeMayo, F. J., Bullock, D. W. &

DiLella A. G. Huang W. M. & Woo, S. L. C. (1988) Luncet 1,497-499 Chirmvin. J. M.. Przvhvla. A. E., MacDonald. R. J. & Rutter, W. J. (1979)

Stoh, U. (1983) Nature 306,332-336

Woo, S. L. C. (1987) Nucktc Ac& Res. 15,1459-1475

Sci. U. 3. A: 85; 757-761 '

Tronche, F., Rollier, A,, Herhomel, P., Bach, I., Cereghini, S., Weiss, M. & Yaniv, M. (1990) Mol. Biol. Med. 7,173-185

Xantho oulos, K. G., Prezioso, V. R., Chen, W. S., Sladek, F. M., Cortese

3811 R. & barnell, J. E., Jr. (1991) Proc. Natl. Acad. Scr. U. S. A. 88, 3807-

Costa R. H. Van-D ke T A Yan, C., Kuo, R., Darnell, J. E., Jr. (1991)

Crossle M. & Brownlee, G. G. (1991) Nature 346,444-446 P d c . N a t l Acad. Jci. 'U. S . 2. 8 7 , 6589-6593

Sakai, $:, Ohtani, N., McGee, T. L., Robhins, P. D. & Dryja, T. P. (1991)

. .

Nature 353,83-86