Proc. Natl. Acad. Sci. USA Vol. 89, pp. 6540-6544, July 1992 Selective loss of a DNase I hypersensitive site upstream of the tyrosine aminotransferase gene in mice homozygous for lethal albino deletions (chromatin structure/gene regulation/liver differentiation) KENNETH S. ZARET*, PATRICE MILOS*, MARIE LIAt, DEEKSHA BALIt, AND SALOME GLUECKSOHN-WAELSCHt *Section of Biochemistry, Box G-J363, Brown University, Providence, RI 02912; and tpepartment of Molecular Genetics, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461 Contributed by Salome G. Waelsch, April 21, 1992

ABSTRACT Several overlapping chromosomal deletis eukaryotes, virtually all regulatory sequences-such as those sning the albino locus in the mouse cause perinatal lethality of promoters, enhancers, and locus control regions-show when homozygous and a block in the trnscriptional Induction DNase I hypersensitivity when the relevant elements are of various unlinked hepatocyte-specific genes. Studies of such active (10). hypersensitivity is thought to be due to lethal albino deletion homozygotes in perinatal stages revealed a altered DNA structures or to specific DNA binding proteins deficiency in the transcriptional inducibility of the tyrosine that perturb or displace a nucleosome, rendering the adjacent aminotrnsferase (TAT) gene by glucocorticolds; yet, glucocor- sequences available for nuclease attack (11, 12). For exam- ticoid receptor and hormone levels were shown to be uected. ple, in vivo and in vitro footprinting studies have shown that To identify a molecular defect underlying thefailure ofi ble binding of the glucocorticoid receptor (13) and a liver- expression, we examine the chromatin structure of the TAT enriched transcription factor (14) disrupt a nucleosome (15) at gene. Whereas in wild-type animals the TAT becomes a DNase I-hypersensitive site 2.5 kilobase pairs (kbp) up- DNase I hypersensitive at birth, such hypersensitivity falls to stream of the rat TAT promoter (16) and that the DNA develop in lethal albino deletion homozygotes. By contrast, the segment at this site can function autonomously as a gluco- deletions do not affect the appearance of three DNase I-hyper- corticoid response element (13). Nuclease-hypersensitive sensitive sites upstream ofthe TAT promoter in the liver, nor do sites have also been detected in hepatocyte chromatin 3.6 and they affect two hypersensitive sites upstream of the expressed 11 kbp upstream ofthe rat TAT promoter, and both of these a-fetoprotein gene. These fings demonstrate that the abnor- sites function independently as enhancer elements in hepa- mality ofchromatin structure identified in lethal albino deletion tocyte-derived cell lines (17). homozygotes occurs on a highly selective basis. Specifically, Assuming that nuclease hypersensitivity reflects the func- normal differentiation ofthe TAT promoter chromatin appears tional state of regulatory sequences, we investigated the to depend directly or indirectly on the action and product of a effects ofthe lethal albino deletions when homozygous on the gene mapping within the deleted region. chromatin structure of the mouse TAT gene. We therefore identified DNase I-hypersensitive sites upstream of the Cell type-specific abnormalities of differentiation at both the mouse TAT gene and examined the appearance of hypersen- molecular and ultrastructural levels are caused by overlap- sitivity during development. Subsequently, a perturbation in ping deletions around the albino locus on chromosome 7 of the formation of one particular hypersensitive site present in the mouse and result in perinatal lethality (1, 2). In particular, normal newborn mice was discovered in littermates homozy- the developmental induction ofunlinked hepatocyte-specific gous for the lethal albino deletions. These results implicate genes, such as that encoding tyrosine aminotransferase specific gene regulatory sequences as direct or indirect (TAT) (3), is affected by the lethal albino deletions, as targets ofa trans-acting gene product normally encoded in the expressed in the deficiency of transcriptional inducibility of chromosomal region missing in mice homozygous for the these genes in the liver at birth (4-7). Normally, the action of lethal albino deletions. endogenous or exogenous glucocorticoid hormone at the time of birth results in inducible expression of TAT and several MATERIALS AND METHODS other liver-specific genes; in albino deletion homozygotes, Animals. Mouse strains carrying the 9H and Cl4CoS lethal the induction ofthese genes by the hormone is blocked (3, 7). albino deletions were bred at the Albert Einstein College of Recent studies have shown that the steady-state levels of Medicine. The deletions are maintained in the heterozygous glucocorticoid receptor mRNA and protein are not affected state since homozygotes die within several hours after birth. in the lethal albino deletions and that the association of the Newborn deletion homozygotes lack eye pigment and are receptor with the heat shock protein hsp90 remains normal thereby distinguished from pigmented heterozygous and (8). Also, no significant differences were found between wild-type homozygous littermates. Livers were removed glucocorticoid hormone levels of normal and homozygous from animals either at 19 days offetal development or shortly albino deletion newborns (8). Therefore, in lethal albino after birth, or from normal adult mice, and used immediately deletion homozygotes the competence ofgenes to respond to for isolation of nuclei and DNase I digestion. Livers of glucocorticoid induction appears to be affected rather than siblings within the same litter were pooled for each analysis elements in the signal transduction pathway itself. after separating deletion homozygotes from normal hetero- Increasing evidence indicates that the configuration of zygous and homozygous littermates. chromatin plays an important role in the mechanism of Isolation of Nuclei and DNase I Treatment. Nuclei were transcriptional activation (9). In the chromatin of higher prepared by either the procedure of Burch and Weintraub (18), as for the samples shown in Figs. 2 and 5, or a The publication costs ofthis article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" Abbreviations: TAT, tyrosine aminotransferase; GRE, glucocorti- in accordance with 18 U.S.C. §1734 solely to indicate this fact. coid response element; FAH, fumarylacetoacetate hydrolase. 6540 Downloaded by guest on September 27, 2021 Genetics: Zaret et at. Proc. Nati. Acad. Sci. USA 89 (1992) 6541

modification (19) of the procedure of Lichtsteiner et al. (20), Maternal Newborn as for the samples shown in Figs. 3 and 4. Nuclei prepared by normal normal c3H/C3H either method were suspended on ice at a concentration of 10 0 + - 0 + n f ' M A260 units/mi. One-fifth of the total sample was lysed imme- diately by adding a stop solution containing SDS, NaCl, and proteinase K as described (21). One-fifth of the total sample was warmed to 370C, further incubated for 3 min, and then lysed as a control for endogenous nuclease activity. The remaining nuclei were warmed to 370C, DNase I was added (see figure legends for final concentrations), and samples were removed for lysis after 30 sec, 1 min, and 3 min. Genomic DNA was purified from lysed samples and sub- jected to Southern blot analysis as described (21). DNA Hybridization Probes. All TAT probes were derived -0.1* 0 V from the plasmid pmTAT-SH3.7 (22), which contains the Sal I/HindIII genomic fragment spanned by probe B (see Fig. 1). Probes A and B were labeled by nick-translation; the 90-bp 1 2 3 4 5 6 7 8 9 10 11 12 13 14 probe C was labeled with [y-32P]ATP and polynucleotide kinase. The I FIG. 2. Lethal albino deletion homozygous mice fail to develop EcoRI/Cla fragment used to probe a-fetopro- DNase I hypersensitivity at the TAT promoter. Genomic DNA tein chromatin has been described (23). samples from nuclease assays of heterozygous C,'h/C14C,,S (normal) maternal, crh/c'h, c'h/c11H, and 311/c31' newborn liver nuclei were RESULTS cleaved with HindII, electrophoresed in agarose gels, transferred to nitrocellulose, and hybridized to radiolabeled EcoRl/Hindlll probe DNase I Hypersensitivity of the Mouse TAT Promoter in A (see Fig. 1); autoradiographs from two experiments are shown. Normal and Lethal Albino Deletion Homozygous Mice. Previ- Lanes 1, 6, and 10 (0), DNA from nuclear samples lysed without ous studies of the rat TAT promoter showed it to be strongly prewarming to 37°C. The absence of subbands in these samples DNase I hypersensitive in hepatocytes (17); we therefore indicates the absence of endogenous nuclease activity during the preparation of liver nuclei. Liver nuclei at 370C (+) were treated as examined the chromatin structure of the mouse TAT pro- follows: lanes 2, 7, and 11, without nuclease for 3 min, as a control moter as a positive control for our assay. Hepatocyte nuclei for endogenous nuclease (all other lanes are with DNase I at 2 were prepared from normal adult heterozygous females and jug/ml); lanes 3, 8, and 12, 30 sec; lanes 4, 9, and 13, 1 min; lane 5, treated with DNase I; the resulting DNA samples were 3 min. Lane 14 (M), genomic DNA cleaved with Hindlll and mixed analyzed by the indirect end-labeling procedure (24). Ge- with picogram amounts of a Sal I/HindIII fragment (upper subband) nomic DNA was digested with HindIII, fractionated by gel and an EcoRI (partial digest)/Hindll fragment (lower subband) from the TAT gene (see Fig. 1). The -0.1 subband in nuclease samples electrophoresis, blotted to nitrocellulose, and hybridized to migrates slightly faster than the latter fragment, demonstrating that the EcoRI/HindIII probe A as schematized in Fig. 1. In the -0.1 hypersensitive site maps at the TAT promoter. The more addition to the expected 5.7-kbp genomic HindIII fragment, slowly migrating subband in lane 3 is due to cross-hybridization to an abundant subfragment appeared due to cleavage at the nonspecific genomic sequences, as verified by various probing TAT promoter at about -0.1 kbp from the transcription start experiments. site (Fig. 2, lanes 3 and 4). The -0.1-kbp subband was seen in all nuclear samples incubated at 37°C, with or without TAT DNase I-hypersensitive sites gave rise to subbands added DNase I (lane 2), but it was absent in samples not that appeared faint in the perinatal liver samples, presumably warmed to 37°C (control DNA; lane 1), demonstrating that because at that developmental stage the liver is primarily a cleavage at the TAT promoter was due to nuclease digestion hematopoietic organ with hepatocyte nuclei representing of chromatin. However, the much faster appearance of the only ~40% of the total number of nuclei (25). Nevertheless, -0.1-kbp subband in the DNase I-treated samples demon- the mouse TAT promoter, which is active at birth, was found strates that the enzyme cleaves specifically at the promoter to be DNase I hypersensitive in normal newborn animals region in chromatin. In conclusion, the mouse TAT pro- (Fig. 2, lanes 7-9). The presence of an internal control marker moter, like the rat TAT promoter, is strongly nuclease on the same gel, which cleaves at an EcoRI site 205 bp sensitive in normal adult hepatocyte chromatin. upstream of the TAT transcription start site, made possible -111~ -79 -4.1 -0.1

- no- St Hi Sl SCSt Ec Ec Hi 1 kbp

=~~B

-iC

FIG. 1. DNase l-hypersensitive site mapping strategy. (Upper) Map ofthe TAT gene showing only the restriction sites (ref. 22 and this study) used to map positions of nuclease hypersensitivity in chromatin; the restriction enzymes used were EcoRI (Ec), HindIll (Hi), Sac I (Sc), Sal I (Sl), and Stu I (St). Solid squares between double lines represent exon sequences (22); wavy line indicates transcribed region. Open arrows and numbers designate distances in kbp between DNase l-hypersensitive sites and the TAT transcription start site. (Lower) Sets oflines indicate mapping strategies used in Figs. 2-4. Restriction sites at the end points ofthe lines labeled A, B, and C indicate boundaries ofthe main restriction digest bands shown in Figs. 2-4; stippled boxes represent regions used as probes. Thin lines beneath labeled lines indicate the extent of DNase I-generated subband DNA fragments visualized by each probe. Downloaded by guest on September 27, 2021 6542 Genetics: Zaret et al. Proc. Natl. Acad. Sci. USA 89 (1992) a precise localization of the -0.1-kbp DNase I-hypersensi- sites might also be found upstream of the mouse promoter. tive site in the TAT promoter. In contrast, the TAT promoter Chromatin samples were digested with Stu I, fractionated by was not hypersensitive in liver nuclei ofC3H/C3H homozygous gel electrophoresis, transferred to nitrocellulose, and hybrid- deletion mice (lanes 11-13), where the promoter is much less ized to the Sac I/Stu I probe C as shown in Fig. 1. The 90-bp active (26). probe C is short enough to allow accurate mapping of To determine whether the deficiency ofhypersensitivity at upstream hypersensitive sites. In addition to the expected the TAT promoter in cJH/c3H was also seen in CJ4coS albino genomic Stu I fragment of =20 kbp, the autoradiograph deletion homozygotes, we performed the experiment shown displayed subbands due to nuclease cleavage in adult hepa- in Fig. 3. The bands in this figure are displayed by hybrid- tocyte chromatin 7.9 and 11 kbp upstream of the TAT ization to probe B (see Fig. 1), which anneals to both transcription start site (Fig. 4, lanes 1-3). Subbands marked subfragments generated by nuclease cleavage of the 5.7-kbp by asterisks in Fig. 4 were present consistently in nuclear HindIII fragment. When the normal adult liver samples samples not warmed to 37°C and therefore are apparently the shown in Fig. 2 were hybridized to probe B, the original result of cross-hybridization to other genomic sequences. -0. 1-kbp subband and a new, more slowly migrating subband Also, a larger probe, which extended from the Stu I site to the became evident (Fig. 3, lanes 2-4); its size was consistent upstream Sal I site, hybridized to these and many other with its end points at the upstream HindIII site and the TAT genomic fragments under stringent hybridization conditions promoter (Fig. 1). Probe B exhibited significant background (data not shown), indicating the presence of a repeated hybridization to other genomic sequences, including a band sequence element. By contrast, the subbands marked by in control DNA that comigrated with the upper -0.1-kbp arrows in Fig. 4 were consistently detected by probe C in subband (lanes 1, 6, and 15). However, the increased inten- nuclear samples warmed to 37°C for 3 min, and they appeared sities of the upper and lower -0.1-kbp subbands in the much faster in the presence of DNase I, indicating that they nuclease-treated samples were concordant, as expected. The were generated by the presence of nuclease-hypersensitive data show that livers ofcl4CoS/cl4Cos newborn mice also lack sites. hypersensitivity at the TAT promoter (lanes 11-14), similar to The -7.9- and -11-kbp hypersensitive sites were evident the C3H/9H newborns. The liver nuclei of newborns used in in normal mice in both late fetal and newborn stages of Figs. 2 and 3 were prepared by different procedures (see hepatocyte development (Fig. 4, lanes 4-6 and 10-12). A Materials and Methods), demonstrating that the lack of subband due to cleavage at -4.1 kbp was also seen in normal hypersensitivity in lethal albino deletion homozygotes is fetal samples but was fainter in newborn and almost unde- independent of the method of sample preparation. tectable in adult samples. All of these hypersensitive sites The analysis ofDNase I hypersensitivity in late fetal stages were present in hepatocyte chromatin of Cl4CoS/C14CoS ho- yielded a striking result: the hepatocytes of both normal and mozygous albino deletion mice at late fetal and newborn c14CoS/cl4CoS fetal mice at day 19 of gestation lacked hyper- stages (lanes 7-9 and 13-15). We conclude that three sites sensitivity at the TAT promoter (Fig. 3, lanes 15-23). Other become hypersensitive upstream ofthe TAT promoter during DNase I-hypersensitive sites were present in these fetal chro- fetal development ofboth normal mice and deletion homozy- matin samples (see below), demonstrating the specificity ofthe gotes but that hypersensitivity at the promoter site develops lack of hypersensitivity at the TAT promoter. We conclude at birth in normal mice and fails to appear in lethal albino that nuclease hypersensitivity at the -0.1-kbp hypersensitive homozygotes. These findings are summarized in Table 1. site normally appears at birth, correlated with the increase in Albino Deletion Homozygotes Exhibit Normal Chromatin TAT . In homozygous deletion mice, hyper- Structure of the a-Fetoprotein Gene. To address further the sensitivity at the -0.1-kbp site fails to develop concomitantly selectivity of the effect of lethal albino deletions on hepato- with a deficiency in TAT transcription. cyte chromatin structure, we examined the nuclease sensi- Upstream TAT Sites Become DNase I Hypersensitive During tivity of two enhancer elements residing upstream of the Fetal Development and Are Not Affected in Albino Deletion a-fetoprotein gene. Previous studies had shown the levels of Homozygotes. The presence of nuclease-hypersensitive sites mRNA for a-fetoprotein to be undiminished in the livers of upstream of the rat TAT promoter (17) suggested that such newborn albino Cl4Cos deletion homozygotes (4). As shown in

Maternal Newborn r a- 4 I 4 normal normal cl4Co ic4C S nlorrTa A ,-S(4ic. :-S - r-%------0 + + W.

-O1L_ - 0.1'L w 1

-0O.1 'i.A,

3 4 . 15 1 1 ' A FIG. 3. DNase I hypersensitivity at the TAT promoter is induced at birth. Analysis ofliver nuclear DNA samples was similar to Fig. 2 except that the DNAs were hybridized to the Sal I/HindIII probe B (see Fig. 1). Maternal samples (lanes 1-5) were the same as shown in Fig. 2. Lanes 1, 6, and 15 (0), DNA from nuclear samples lysed without prewarming to 37°C. Liver nuclei warmed to 37°C (+) were treated as follows: lanes 2, 7, 11, 16, and 20, without added DNase I for 3 min; lanes 8-10 and 12-14, with DNase I at 1 ug/ml; lanes 17-19 and 21-23, with DNase I at 3 jig/ml for 30 sec (lanes 8, 12, 17, and 21), 1 min (lanes 9, 13, 18, and 22), or 3 min (lanes 10, 14, 19, and 23). The probe detects both subbands generated by DNase I cleavage at the TAT promoter (-0.1; open arrows). Other bands present are due to cross-hybridization of the probe to nonspecific genomic sequences. Downloaded by guest on September 27, 2021 Genetics: Zaret et al. Proc. Natl. Acad. Sci. USA 89 (1992) 6543

Maternal Newborn Fetal Normal c3H/c3H normal nor. &4CoS/c14cos nor. cd4CS/cl4c.S 0 + _-- - 0 + + - + o + - 4 - - ---_ _AAM~ 4 i~~ , -7.9 JOEI . A, :i f pw * In .... -.... IEIE He 6; g * g g A t A <'-4.1

!:;,i-,.',', .'' 9"$%1: -. A w .....; A.... 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 FIG. 5. Hypersensitive sites upstream of the a-fetoprotein gene are not affected in lethal albino deletion homozygotes. Analysis of FIG. 4. Upstream TAT hypersensitive sites are present in fetal the newborn liver nuclear DNA samples was the same as for that development. Liver nuclear DNA samples were cleaved with Stu I shown in Fig. 2, except that the DNAs were digested with EcoRI and and analyzed as in Fig. 2 except that they were hybridized to the hybridized to a 1-kbp EcoRI/CIa I fragment upstream of the a-fe- radiolabeled Sac I/Stu I probe C (Fig. 1). The newborn and fetal toprotein gene (23). Open arrows show the positions ofsubbands due samples were taken from the DNase I-treated samples shown in Fig. to cleavage at the upstream enhancers El and ElI. The lane-to-lane 3, whereas the maternal sample was treated differently. Lane 1, variation in intensity of the subbands was also seen in a previous maternal nuclear sample lysed without prewarming to 370C; lanes 2 study (23). The subbands below the ElI subbands were not seen and 3, nuclei at 37C (+) in the absence of added nuclease for 5 min consistently. (lane 2) or in the presence of DNase I at 3 jtg/ml for 30 sec (lane 3). The probe detects subbands (open arrows) generated by DNase I cleavage at -11, -7.9, and -4.1 kbp with respect to the TAT appearance of these sites. In contrast, the chromatin of the transcription start site. In some experiments, DNase I hypersensi- TAT promoter itself was not nuclease hypersensitive in late tivity was observed at a site between the -7.9- and -11-kbp sites. fetal stages of normal mice, but hypersensitivity appeared in Probe C often gave rise to bands (asterisks) that were visible in perinatal stages. Thus, in the case of the TAT promoter, as nuclease samples and samples that were lysed without warming to shown for numerous other inducible regulatory elements 37rC (data not shown); these are likely to be due to hybridization to (10), the appearance of nuclease hypersensitivity coincides nonspecific genomic sequences. with increased gene activity. Strikingly, both the appearance 5 of TAT promoter hypersensitivity and transcriptional induc- Fig. (lanes 1-4), normal newborn mice exhibit nuclease ibility are missing in albino deletion homozygotes. The pres- hypersensitivity at the EI and EII a-fetoprotein enhancers ent results indicate that the failure of TAT inducibility in (27), similar to that observed in an earlier study (27, 28). deletion homozygotes is not due to a complete perturbation Importantly, both enhancers are also hypersensitive in the of TAT chromatin structure. This conclusion is consistent C3H/c3H deletion homozygous newborn mice (lanes 5-8). The with the fact that TAT transcription does in fact continue in albino deletions therefore affect liver chromatin with a high the mutants even though at a low basal level (3, 5, 6, 26). Also, degree of selectivity, perturbing the formation of a hyper- hypersensitive sites at enhancers of the a-fetoprotein gene sensitive site only at the TAT promoter and not at five other were shown to be unaffected in the deletion homozygotes, hypersensitive sites tested here. further demonstrating the specificity of the effect found in TAT chromatin. Moreover, the perinatal development of DISCUSSION nuclease hypersensitivity at the TAT promoter in normal The studies reported here concern further molecular inves- animals appears to identify a change in chromatin structure tigations ofthe failure ofthe TAT gene to be transcriptionally critical for normal transcriptional induction of the TAT gene induced at birth in c3H and cl4cos albino deletion homozy- at birth. All of this leads to the suggestion that a factor(s), gotes. This system serves as a model for identification of possibly gene products normally encoded in the deleted possible molecular defects associated with the failure of a region and deficient in the deletion homozygotes, plays an specific cluster of hepatocyte-specific genes to show normal essential role in a chain of events leading to this change of inducible expression in newborn deletion homozygotes. Nu- chromatin structure. clease-hypersensitive sites were detected 4.1, 7.9, and 11 kbp When speculating about the possible mechanisms under- upstream of the TAT promoter in the chromatin of normal lying the normal perinatal changes in the TAT promoter hepatocytes. The three sites were hypersensitive in nuclei of chromatin structure, it is important to note that this promoter fetuses prior to the induction of TAT transcription, and the had acquired hypersensitivity to both exogenous DNase I and homozygous lethal albino deletions had no effect on the an endogenous nuclease, indicating that the DNA had been Table 1. Summary of hypersensitive site analysis Nuclease-hypersensitive site Developmental stage Genotype Phenotype -11 kbp -7.9 kbp -4.1 kbp -0.1 kbp Adult +/- Normal + + ± + Newborn +/+ and +/- Normal + + + + Fetal +/+ and +/- Normal + + + - Newborn -/- Mutant + + + - Fetal -/- Mutant + + + - Downloaded by guest on September 27, 2021 6544 Genetics: Zaret et al. Proc. Natd. Acad. Sci. USA 89 (1992) exposed to distinct probes. A comparison with other systems the deletion homozygotes. Molecular studies of additional (10, 11, 12) suggests that such nuclease hypersensitivity may affected genes, such as those encoding phosphoenolpyruvate reflect a disruption or loss of local nucleosome structure, carboxykinase and serine dehydratase, as well as the precise presumably due to the binding of trans-acting factors. The identification ofthe relevant regulatory gene(s) deleted in the absence of such nuclear factors in the albino deletion ho- 9H and~4CoS homozygous mice, are required in order to mozygotes, and therefore their failure to alter the chromatin provide further information about the cascade of events structure of the TAT promoter, may be responsible for the involved in terminal hepatocyte differentiation that are af- inability of the promoter to become inducible at birth. fected in these mutants. Previous studies showed that administration of glucocor- ticoids to albino deletion homozygous newborns did not We thank Drs. Gfnther Schfitz and Shirley Tiighman for gener- enhance the inducibility ofTAT transcription (3) but that the ously providing DNAs for probes. We are particularly grateful to Dr. signal transduction pathway for the steroid Ronald DePinho for many discussions and his generous help with the appeared to be manuscript. Finally, we thank Vivian Gradus for her untiring help unaffected (8). In the rat, a glucocorticoid response element and her skills in preparing this manuscript. This work was supported (GRE) resides 2.5 kbp upstream of the TAT promoter, and by grantsfrom the National Institutes ofHealth to K.S.Z. (GM36477) the sequence becomes DNase I hypersensitive in mature and S.G.-W. (GM27250) and by Grant CD-38 from the American hepatocytes as a result of hormone administration (16). Cancer Society to S.G.-W. Detailed studies have shown that hormone stimulation causes the glucocorticoid and additional 1. Gluecksohn-Waelsch, S. (1979) Cell 16, 225-237. receptor binding factors 2. Gluecksohn-Waelsch, S. (1987) Trends Genet. 3, 123-127. (13, 14) to disrupt a nucleosome at the -2.5-kbp site (15), 3. Schmid, W., Mfiller, G. & Gluecksohn-Waelsch, S. (1985) resulting in an active GRE (29). The rat TAT promoter does Proc. Nat!. Acad. Sci. USA 82, 2866-2869. not appear to bind the glucocorticoid receptor, and it requires 4. Sala-Trepat, J. M., Poiret, M., Sellem, C. H., Bessada, R., the upstream GRE for hormonal regulation (13). Since the Erdos, T. & Gluecksohn-Waelsch, S. (1985) Proc. Nat!. Acad. mouse and rat TAT promoter sequences are virtually iden- Sci. USA 82, 2442-2446. tical (22), it seems unlikely that hypersensitivity at the mouse 5. Loose, D. S., Shaw, P. A., Krauter, K. S., Robinson, C., TAT promoter at birth would be due to direct binding of the Englard, S., Hanson, R. W. & Gluecksohn-Waelsch, S. (1986) glucocorticoid receptor. Rather, nuclease hypersensitivity Proc. Nat!. Acad. Sci. USA 83, 5184-5188. might reflect binding of 6. Donner, M. E., Leonard, C. M. & Gluecksohn-Waelsch, S. the receptor to a site upstream, and (1988) Proc. Nat!. Acad. Sci. USA 85, 3049-3051. a regulatory gene product deleted in the cH and C14CoS 7. DeFranco, D., Morris, S. M., Jr., Leonard, C. M. & Glueck- homozygous mice might be required for the purpose of sohn-Waelsch, S. (1988) Proc. Nat!. Acad. Sci. USA 85, making the promoter competent to respond to the inducing 1161-1164. signal. Another possibility is that hypersensitivity at the 8. DeFranco, D., Bali, D., Torres, R., DePinho, R. A., Erickson, promoter could be a consequence of the regulatory gene R. P. & Gluecksohn-Waelsch, S. (1991) Proc. Nat!. Acad. Sci. product acting directly at a distal upstream site. A thorough USA 88, 5607-5610. understanding of TAT gene regulation awaits a developmen- 9. Felsenfeld, G. (1992) Nature (London) 35S, 219-224. tal study and a definition of the glucocorticoid-responsive 10. Gross, D. S. & Garrard, W. T. (1988) Annu. Rev. Biochem. 57, 159-197. sequences in the mouse. The chromatin abnormality identi- 11. McGhee, J. D., Wood, W. I., Dolan, M., Engel, J. D. & fied in a gene affected by the albino deletions represents at Felsenfeld, G. (1981) Cell 27, 45-55. this time the strongest evidence for the existence of a 12. Benezra, R., Cantor, C. R. & Axel, R. (1986) Cell 44, 697-704. trans-acting factor essential for normal differentiation of the 13. Jantzen, H.-M., Strnhle, U., Gloss, B., Stewart, F., Schmid, affected target genes. The analysis of a molecular defect W., Boshart, M., Miksicek, R. & Schutz, G. (1987) Cell 49, associated with the abnormalities of TAT expression and 29-38. inducibility reported here raises the question of identical 14. Rigaud, G., Roux, J., Pictet, R. & Grange, T. (1991) Cell 67, effects ofthe deletions on the chromatin ofthe other similarly 977-986. affected genes encoding gluconeogenic liver enzymes (1, 6). 15. Reik, A., Schfitz, G. & Stewart, A. F. (1991) EMBO J. 10, have 2569-2576. Recent chromosome mapping experiments (30) 16. Becker, P. R., Renkawitz, R. & Schuitz, G. (1984) EMBO J. 3, shown that the gene encoding fumarylacetoacetate hydrolase 2015-2020. (FAH) is included in the lethal albino deletions, which, in 17. Nitsch, D., Stewart, A. F., Boshart, M., Mestril, R., Weih, F. addition to their effects on inducible expression of certain & Schuitz, G. (1990) Mol. Cell. Biol. 10, 3334-3342. liver genes, had been shown earlier to cause abnormalities of 18. Burch, J. B. E. & Weintraub, H. (1983) Cell 33, 65-76. the ultrastructure of particular organelles of hepatocytes and 19. Mirkovitch, J. & Darnell, J. E., Jr. (1991) Genes Dev. 5, 83-93. kidney tubules (1). FAH catalyzes the terminal step in 20. Lichsteiner, S., Wuarin, J. & Schibler, U. (1987) Cell 51, tyrosine catabolism, and a deficiency in this enzyme in 963-973. humans has been reported to cause tyrosinemia type 1 (31). 21. Liu, J.-K., Bergman, Y. & Zaret, K. S. (1988) Genes Dev. 2, 528-541. The disease is marked by accumulation of a toxic metabolic 22. Muller, G., Scherer, G., Zentgraf, H., Ruppert, S., Herrmann, derivative oftyrosine, depletion ofglutathione, and inhibition B., Lehrach, H. & Schfitz, G. (1985)J. Mol. Biol. 184, 367-373. of S-adenosylmethionine synthetase (31). Similar metabolic 23. Godbout, R. & Tilghman, S. M. (1988) Genes Dev. 2, 949-956. abnormalities caused by the absence of FAH might be 24. Wu, C. (1980) Nature (London) 286, 854-60. suspected to be responsible for some of the effects observed 25. Paul, J., Conkie, D. & Freshney, R. I. (1969) Cell Tissue Kinet. in the lethal albino deletion homozygous mice. However, the 2, 283-294. high degree of specificity of the defects in fetal and newborn 26. Tonjes, R. R., Xanthopoulos, K. G., Darnell, J. E., Jr., & livers of lethal albino deletion homozygotes, on both ultra- Paul, D. (1992) EMBO J. 11, 127-133. structural and biochemical levels (1, 2), makes it appear very 27. Godbout, R., Ingram, R. & Tilghman, S. M. (1986) Mol. Cell. Biol. 6, 477-487. unlikely that FAH enzyme deficiency itself could cause the 28. Nahon, J.-L., Venetianer, A. & Sala-Trepat, J. M. (1987) Proc. entire spectrum of abnormalities. In this connection, it ap- Nat!. Acad. Sci. USA 84, 2135-2139. pears particularly significant that the effect on the chromatin 29. Richard-Foy, H. & Hager, G. L. (1987) EMBO J. 6, 2321-2328. of the TAT promoter described here is also highly specific, 30. Klebig, M. L., Russell, L. B. & Rinchik, E. M. (1992) Proc. since other nuclease-hypersensitive sites in TAT and the Nat!. Acad. Sci. USA 89, 1363-1367. a-fetoprotein gene (Figs. 4 and 5, Table 1) are not affected in 31. Kvittingen, E. A. (1986) Scand. J. Clin. Lab. Invest. 46,27-34. Downloaded by guest on September 27, 2021