Proc. Natl. Acad. Sci. USA Vol. 89, pp. 6540-6544, July 1992 Genetics 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). Nuclease 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 promoter 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.