expression of the gene and various deleted constructions. In addition, exon 6 .... The reaction was stopped by the addition ofan equal volume ..... schematic representation ofK18 is shown with exons represented by ..... The top portion shows.
MOLECULAR AND CELLULAR BIOLOGY, Mar. 1993, P. 1815-1823
Vol. 13, No. 3
0270-7306/93/031815-09$02.00/0 Copyright © 1993, American Society for Microbiology
cis Regulation of the Keratin 18 Gene in Transgenic Mice NICKOLAY S. NEZNANOV
AND
ROBERT G. OSHIMA*
Cancer Research Center, La Jolla Cancer Research Foundation, 10901 North Torrey Pines Road, La Jolla, California 92037 Received 24 August 1992/Returned for modification 27 November 1992/Accepted 17 December 1992
The gene coding for human keratin 18 (K18), a type I intermediate filament protein found in a variety of simple epithelia, is regulated correctly in transgenic mice but is promiscuously expressed after direct transfection into cell culture lines. We have begun an investigation of the mechanisms responsible for the correct regulation of K18 with a comparison of the chromatin state of K18 in permissive and nonpermissive transgenic mouse tissues to identify seven expression-specific, DNase-hypersensitive sites that correlate with known or potential regulatory regions of the gene. Four of these sites are associated with the proximal promoter region and the first intron that has been implicated previously in the transcriptional control of K18. Two hypersensitive sites are associated with a conserved Alu repetitive sequence located immediately upstream of the proximal promoter elements. Transcription of this Alu element in a direction opposite that of K18 was correlated with K18 expression in transgenic tissues. The final hypersensitive site was mapped to exon 6. The potential importance of this region for the expression of K18 was supported by the results of transient expression of the gene and various deleted constructions. In addition, exon 6 and the intron 1 regulatory region were distinguished from the remainder of K18 by differential DNA methylation in expressing and nonexpressing tissues. The CpG-rich proximal promoter and first exon regions remain unmethylated in both permissive and nonpermissive tissues. These results suggest that DNA methylation is not the primary mechanism of control of the gene. An Alu RNA polymerase III transcription unit and exon 6 are implicated in regulation of K18. The tissue-specific expression of many eukaryotic genes is controlled by the interaction of particular transcription factors with their cognate binding sites and with additional components of the transcription complex (12). Genes which retain their original cell type specificity after reintroduction into appropriate differentiated cell types provide the opportunity of defining both the sequence elements and the trans-acting factors that may restrict expression. However, some tissue-specific genes are expressed inappropriately after transfection into cells even though the same genes are expressed correctly in transgenic mice (1, 8, 15, 17, 18, 34, 37). In some of these cases inappropriate expression in cell cultures is observed even when only a few gene copies are introduced. These results indicate that titration of inhibitory activities is an unlikely explanation of the promiscuous expression (18, 37). The apparent paradox of correct regulation in transgenic mice and promiscuous expression after direct transfection in culture may be due to a difference in the chromatin state of the genes. The chromatin structure of the endogenous genes in nonpermissive tissues may restrict accessibility to necessary transcription factors. Several intermediate filament genes may be regulated in this way (24). The human keratin 18 (K18) gene is a particularly clear example of this type of regulation. K18 is the ancestral intermediate filament gene from which more specialized type I keratins have evolved (5). The mouse homolog of the K18 protein (Endo B or mK18) copolymerizes with the complementary and coexpressed type II murine keratin 8 (mK8 or Endo A) (6, 10, 22, 23) to form the first intermediate filaments expressed during mouse development (14, 26). In murine embryonal carcinoma cells, expression of mouse and human K18 genes appears to be limited, at least in part, by low levels of AP1 transcription factor activity. The combination of jun and fos gene products *
that compose AP1 activity acts through a complex transcriptional enhancer region located within the first intron of the gene (25). In adults, K18 expression is limited, with few exceptions, to a variety of simple epithelial cell types, including those in intestine, liver, and kidney. K18 expression is not found in muscle or spleen and is severely restricted in brain to the ependymal simple epithelium (1, 11, 21). Fusion of human and mouse differentiated cells which differ with regard to K18 expression results in persistent expression of K18, which was active, and no change in the mK18 gene, which was silent (28). This indicates the repressed state of the silent mK18 gene is very stable and due to cis-acting mechanisms. Promiscuous expression of K18 is observed after direct transfection into fibroblasts or myoblasts that do not express endogenous mK18 (18, 21a). In contrast, in transgenic mice K18 is regulated entirely appropriately with regard to both tissue specificity and the efficiency of expression (1). We have examined the chromatin and DNA methylation state of K18 in transgenic mice. These studies reveal several potential regulatory regions of K18 which are in different chromatin states in permissive and nonpermissive tissues. Included among these regions are a putative RNA polymerase III transcription unit that is regulated in concert with K18 and an internal regulatory element located within a coding exon. Furthermore, DNA methylation does not appear to account for the repressed chromatin state in nonpermissive mouse tissues. MATERIALS AND METHODS
Identification of DNase hypersensitive sites (HSS). Nuclei from various tissues of K18 transgenic mice by Wu (35). Nuclei were suspended in nuclear buffer {60 mM KCl, 15 mM NaCl, 5 mM MgCl2, 0.1 mM EGTA [ethylene glycol-bis (P-aminoethyl ether)N,N,N',N'-tetraacetic acid], 15 mM Tris-HCl [pH 7.4], 0.5 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, were purified as described
Corresponding author. 1815
1816
NEZNANOV AND OSHIMA
300 mM sucrose, 5% glycerol}, supplemented with 0.1 mM CaCl2 at a concentration equivalent to 200 ,ug/ml of DNA. DNase I was added to 0.5 to 6 ,ug/ml, and the mixture was incubated at 30°C for 10 min. Nuclei from liver were digested by endogenous DNase for 3 to 8 min at 30°C without additional nuclease supplementation in some experiments. The reaction was stopped by the addition of an equal volume of 1% sodium dodecyl sulfate (SDS)-20 mM EDTA. After digestion with proteinase K (200 ,ug/ml) for 2 to 3 h at 37°C and extraction with phenol:chloroform, the DNA was recovered by ethanol precipitation. DNAs were digested with restriction enzyme (7 U/,ug of DNA) for 4 h at 37°C. Aliquots (10 ,g) of cleaved DNA were separated by electrophoresis in 0.8% agarose and transferred to charged nylon filters (Zeta-probe; Bio-Rad) in 0.4 M NaOH. DNA was cross-linked to the membranes by exposure to 1,200 J of UV light per m2 (Stratagene, San Diego, Calif.). Probe I (K18 gene sequence; GenBank accession number M19353; nucleotides [nt] 456 to 810) (19) was generated by 10 cycles of polymerase chain reaction (PCR) with one primer, 32P-dCTP, and 100 ng of previously purified fragment. Probe V, corresponding to the first exon of K18, was made by SP6 RNA polymerase transcription of the BglII-digested pK189 plasmid (27). Probes II (K18 nt 4017 to 4934), III (nt 4934 to 5640), IV (nt 5308 to 6309), and IX (nt 3414-4017) were prepared by the random primer method (9) using 32P-dClTP and 25 ng of fragments isolated by appropriate restriction enzyme digestion and gel purification. Hybridizations were performed at 60°C in 5x SSPE (lx SSPE is 0.18 M NaCl, 10 mM NaPO4, and 1 mM EDTA [pH 7.7]) with final washes at 650C in 0.2x SSPE-0.1% SDS. Membranes were stripped of probes by incubation at 100°C for 15 min in 0.1% SDS. Transfection and RNA analysis. NIH 3T3 and HR9 mouse endodermal cells were transfected by the calcium phosphate precipitate method (31) with 20 p,g of DNA per 9-cm dish of cells. All plasmids were cotransfected with 2 p,g of the pMClNeopA plasmid (33) to normalize the transfection efficiency. RNA was purified by acidic phenol extraction of cells lysed in 0.5% SDS-20 mM EDTA (31). Total RNA was treated with RNase-free DNase I at 37°C for 60 min in the presence of RNase inhibitor (Stratagene). K18 and NeoR RNAs were measured by RNase protection analysis (20) using 32P-UTP labeled probes. The RNA probe for K18 RNA was the result of T7 RNA polymerase transcription of the EcoRI-digested BSK18rp plasmid. This plasmid contains a 431-bp fragment (nt 2284 to 2715) of the K18 gene overlapping the RNA start site as reviously detected by S1 nuclease analysis (18). The Neo RNA was detected with a SP6 RNA polymerase transcript of the EcoRI-digested NeoSP6rp plasmid, which represents a 245-bp EcoRI-toNarI fragment of pMClneopA subcloned first into Bluescript KS and then into the pGEM1 vector via a fragment generated by EcoRI and XhoI. Both probes were added together to transfected cell RNAs for hybridization at 43°C. Protected probe was revealed by digestion with RNases A and Ti and analysis by acrylamide gel electrophoresis in 8 M urea and autoradiography. Detection of the transcripts of the Alu sequence proximal to the K18 promoter was performed with probes synthesized by SP6 polymerase from an EcoRI-digested plasmid designated ExolOGem, containing a 623-bp fragment of the K18 gene (nt 1952 to 2575). Exon 6 constructions. The 10-kb K18 gene is carried on the pGEM1 plasmid (18). Substitution of K18 exons 2 to 7 for the corresponding portion of the cDNA was facilitated by the
MOL. CELL. BIOL.
unique BglII site located at the beginning of exon 2 and the unique KpnI site located within the noncoding region of exon 7, upstream of the polyadenylation signal. Plasmid constructions 1 and 3 to 6 were generated by closing the appropriately digested plasmids, followed by substitution into the gene in some cases. For plasmids 7 and 8, exon 6 was amplified by PCR and cloned with XbaI linkers into the BSKSM13+ vector. A NotI-KpnI fragment containing exon 6 was then cloned between either the BglII and KpnI sites (construction 8) or the BamHI and KpnI sites (construction 7) of plasmid 2. Construction 9, in which 33 bp of exon 6 were replaced with corresponding coding sequences of the K19 cDNA, was made by introducing a synthetic double-stranded oligonucleotide between the BamHI and XmaI (SmaI isoschizomer) sites of exon 6 in a fragment of the whole K18 gene. Subsequently a fragment defined by a unique BstEII site in exon 5 and a Kpnl site in exon 7 was used to replace the wild-type K18 gene fragment. A larger replacement of exon 6 in construction 10 was accomplished by introducing a MstII site into the 5' end of a portion of the K19 cDNA homologous to K18 exon 6 by use of synthetic oligonucleotides and PCR. In addition, an XhoI site was introduced by the same method into the K18 sequence near the 3' end of exon 6. The K19 MstII-XhoI fragment was introduced into a K18 gene fragment and finally into the whole gene. The coding potential of constructions 9 and 10 was confirmed by transfection into mouse HR9 cells and immunofluorescent staining of K18 incorporated into the endogenous mouse keratin filament network. RESULTS DNA methylation state of the K18 transgene. The first exon and proximal promoter regions of both the human and mouse K18 genes are G+C rich and contain an unusual abundance of the normally underrepresented CpG dinucleotide (Fig. 1D). These regions resemble CpG islands (3, 4) which are distinguished by their lack of DNA methylation and their common association with constitutively active promoters. Previous studies of mK18 in cultured differentiated cell lines had revealed that this region was highly methylated in cell lines that did not express the gene and not methylated in permissive cell lines (28). Therefore we have investigated the methylation state of K18 in three lines of transgenic mice, all of which express the transgene efficiently and in appropriate tissues. Figure 2 shows the result of Southern blot analysis of K18 transgenic mouse DNAs that have been digested with the methylation-sensitive restriction enzymes HpaII or the isoschizomer MspI. MspI cleaves the CCGG recognition sequence regardless of the methylation state of the internal CpG dinucleotide, while HpaII cleaves the sequence only if it is not methylated. The endogenous coding K18 gene (mK18 1-1) and the inactive pseudogene (mK18 1-2) provide supporting controls for the state of the K18 transgene. The mK18 1-2 pseudogene is completely resistant to digestion with HpaII in all tissues. The absence of the mK18 1-2 band in samples digested with MspI (lanes 2, 6, and 11) confirms that CCGG recognition sequences are present in the mK18 3-2 fragment. The first exons of both K18 and mK18 1-1 are digested by HpaII to fragments too small to detect in all tissues tested. Thus in contrast to fibroblasts and myoblast cell lines, the first exon of mK18 1-1 is not substantially methylated in nonexpressing tissues (skeletal muscle, heart, brain). It is likely that the methylation previously observed in cell lines is a consequence of the de novo methylation activity reported for several types of cells in culture (2).
A
1817
cis REGULATION OF K18 GENE
VOL. 13, 1993
K18 Gene
1
r,i
4 4
2
3
n
m
1
4 5 m
7
6
..-_-
-E*
S
4.4
-
2.3
-
1
TG1 r Li Li Br Li
Li
M H
H
-
M
2
4
5
6
3
TG3 TG2 1 r- [ Li Mu He Li Li LI Mu Br H H 7 8
H
-
2.0 -
100
oII
__ _t
M
9 10 11
i
B % methylated Liver
Li
a.
H 12
H H 13 14
-
11
-
K18
i.
I_ .1
Spleen
E32 100 0.5 -
C Hpall + Hhal
D
E
CG dinucleotides
GC dinucleotides
FIG. 1. Summary of the DNase HSS and methylation state of K18 in transgenic mice. (A) The K18 exons are numbered and depicted as filled boxes. The open boxes indicate the positions of two Alu repetitive elements. DNase HSS are indicated by the arrows. (B) Summary of the methylation state of HpaII and HhaI sites within K18. Note that sites present in the first exon are not methylated while additional sites within exons 3, 4, and 7 are completely methylated. Sites located within the first intron and exon 6 are less methylated in liver than in spleen. The two sites located within exon 6 have been separated for clarity but are actually separated by only 17 bp. (C) All HpaII and HhaI sites are shown. (D) Location of CG dinucleotides. Note the clustering of this dinucleotide in the proximal promoter and first exon of the K18 gene. (E) GC dinucleotide positions for comparison to the lessfrequent use of CG dinucleotides.
Differential methylation of this region is not the cause of the transcriptional silence of these genes. The methylation states of additional portions of K18 were assessed by the use of different probes and the use of an additional methylation-sensitive restriction enzyme, HhaI (Fig. 3). The methylation state of DNA from liver, which expresses K18, was compared with that of DNA from spleen and brain, which do not substantially express the transgene. The first HpaII site immediately downstream of the first exon was not highly methylated (Fig. 3B, lanes 3 to 5). However, the following HhaI site was found to be differentially methylated depending on the tissue of origin (Fig. 3F, lanes 2 to 4; compare the relative intensities of bands h and i in each lane). DNA from transgenic liver was digested primarily to fragment i (Fig. 3F, lane 2) (the differences in absolute intensities between lanes is due to unequal loading), while digestion of spleen DNA resulted primarily in fragment h (Fig. 3F, lane 4). In contrast to sites within the first exon, two HpaII sites located within exons 4 and 7 and one HhaI site located within intron 2 were completely methylated in liver, brain, and spleen DNAs (Fig. 3C, lanes 3 to 5; Fig. 3D,
FIG. 2. DNA methylation state of the first exon of the transgenic human K18 gene and the endogenous mouse homolog, mK18 p-1. DNAs from three lines of K18 transgenic mice (TG1, TG2, and TG3) were analyzed by Southern blot analysis using an RNA probe representative of the first exon of mK18 1-1 which also recognizes K18. DNAs from liver (Li), brain (Br), skeletal muscle (Mu) or heart (He) were digested with PstI and EcoRI and as indicated with MspI (M) or HpaII (H). Radioactive size markers are indicated at the left in kilobases. Identities of the endogenous coding mK18 1-1 (1-1), the mK18 1-2 pseudogene (,B-2), and K18 are indicated on the right. Note that mK18 13-2 pseudogene is completely resistant to digestion with HpaII while K18 and mK18 1-1 are digested to fragments too small to detect in all tissues. The same filter was also hybridized with a hK18 probe that confirms the assignment of the transgene (data not shown).
lanes 3 to 5; and Fig. 3F, lanes 2 to 4). One HpaII site located within exon 6 was differentially methylated when DNAs from expressing and nonexpressing tissues were compared (Fig. 3E, lanes 3 to 5; compare the intensities of fragments f and g). In addition, a neighboring HhaI site 17 bp downstream in exon 6 was also differentially methylated when liver DNA was compared to either brain or spleen DNA (Fig. 3G, lanes 2 to 4; compare fragments f and k within each lane.) In additional experiments, all sites discussed were tested in each of the three K18 transgenic lines with very similar results (data not shown). After estimation of the degree of digestion of particular sites by densitometry, the deduced modification of each site was averaged for all three transgenic lines. The results are shown in Fig. 1B. Three different patterns of methylation are apparent. First, the proximal promoter and first exon, which are CpG rich, are not methylated in any tissue. In contrast, certain sites within the body of the gene are completely methylated in all tissues analyzed. Finally sites located within intron 1 and exon 6 are highly methylated in spleen, which does not express K18, and less methylated in liver, which expresses the gene. DNase HSS of the K18 transgene. To locate potential regulatory regions, the chromatin form of the K18 transgene in different tissues was subjected to limited digestion with DNase I. In an actively transcribed gene, DNase HSS commonly coincide with regulatory regions which bind transcription factors and induce sensitivity to nuclease digestion as well as prevent protective nucleosome formation (12). DNAs from DNase I-treated liver and spleen nuclei of K18TG1 mice were isolated and analyzed by Southern
NEZNANOV AND OSHIMA
1818
MOL. CELL. BIOL.
A. 11
DC
Probos:
NI
A
X
_ _K= 1IIIIIIII=
B
aKb
ion
1 1IIIl11
_
Ap Bwa
SH
K Barn
Bam
Sp5een
1 234 5567
456
Hhal1
GRP78 Liver
Spleen ffi
W6
481
1011
K18 Liv2
8 9
s
-23 _c b
a
-9.4
d
-6.6
e
k
--
-4.4
........ .:
Probes:
B.
C.
D.
IX
III
X
E. x
F.
G.
11
III
0__ve__ ce_
23-
23
94 8f -
-.SMt
--O
-2.3 -2.0 --1.2
-
964-
X
23-
94-
44 -
t6
4.4 -
I
-
4.4
23 20-
2 3a
2.0-
-b
40
*,*a
a-
2 3-
I
-h
20 0_
* -g
*-c
a
go
h-e
0 56 -
.
-k
056-
0.56-
0
1 2 3 4 5 Sae Kpn
2 3 45 1 2 3 4 5 1 2 3 4 5
BH
m
Hi
1 2 3 4
2 3 4
SIDC I.+ Kpn Hhal
"I
Lig
v
88. Hpo Hpa 11 FIG. 3. DNA methylation state of K18 in transgenic mice. (A) A schematic representation of K18 is shown with exons represented by filled boxes. Open boxes represent two Alu type repetitive elements located upstream of the first exon. Portions of the gene used as probes are shown above the gene with the solid or open lines and designated with roman numerals. Restriction enzyme sites for ApaI (Ap), BamHI (Bam), Sacl (S), HincIl (H), and KpnI (K) are indicated on the map below the gene. HhaI and HpaII sites are indicated above and below the map line. Fragments detected in the analysis are indicated by the double-headed arrows and lowercase letters (a to i). Southern blot analyses of the transgenic K18TG1 line (B to E) and K18TG3 line (F and G) are shown. (B, C, and D) Exposures of the same filter analyzed sequentially with probes IX, III, and X, respectively. K18TG1 DNA was digested with BamHI and additionally as indicated below panel B with MspI or HpaII. Lanes 1, 2, and 3 contained liver DNA, and lanes 4 and 5 contained brain and spleen DNA, respectively. Note in panel B that fragment a is the product of HpaII digestion of all three tissues (lanes 3 to 5), indicating that the HpaII site which defines the 5' end of the fragment was completely cut and thus not methylated. In panel E each lane contained the same DNAs as for panels B to D; however, the DNA was digested with Sacl and KpnI instead of BamHI. Note the difference in intensity of fragments f and g in panel E (lanes 3 and 5), which indicates different degrees of digestion of a HpaII site within exon 6. In panels F and G, K18TG3 mouse DNA was used. Lanes 1 and 2 contained liver DNA, while lanes 3 and 4 contained brain and spleen DNA, respectively. The samples were digested with Sacl and KpnI and HhaI as indicated below panel F. Fragments marked with stars in panel F were likely generated by DNase HSS within the proximal promoter and intron 1 (see Fig. 5) during dissection. Note the differences in the relative intensities of fragments h and i in panel F and fragments f and k in panel G, lanes .
. .
11
.
.
-
2 to 4.
blotting. The results of hybridization with a K18 probe and a control gene probe for the glucose-regulated protein (GRP78) (29) are shown in Fig. 4. In panel A, the K18specific band marked c+e appears after nuclease digestion of liver but not spleen nuclei. The appearance of the band correlates with expression of K18 in liver but not in spleen (1). A control, constitutively active gene (GRP78) contains a
Pst I EcoRI FIG. 4. Tissue-specific DNase HSS of K18 in transgenic mice. (A) Liver and spleen nuclei from K18TG1 mice were treated with increasing amounts of DNase I. DNAs were purified, digested with PstI, and subjected to Southern blot hybridization analysis in panel A by using a K18 probe for the first exon (probe V, Fig. 5). The arrow marker c+e indicates the position of fragments generated by HSS within K18 in liver but not in spleen chromatin. (B) The same DNAs were digested with EcoRI and analyzed with a cDNA probe for the mouse glucose regulated protein (GRP78). In this case the single GRP78 gene fragment of 5.9 kb is digested to one or two fragments indicated by arrows. In contrast to HSS of K18, which are only found in liver, GRP78 HSS are detected in both spleen and liver.
particularly sensitive site in spleen as well as liver nuclei (upper arrow, panel B), as well as an additional weaker site in liver (panel B, lower arrow). Thus, the lack of K18 HSS in spleen is not due to insufficient nuclease digestion. Indeed, the strong GRP78 HSS in spleen nuclei is even revealed by the action of endogenous DNase (Fig. 4B, lane 6). The c+e band of K18 in liver was subsequently found to be the combination of products generated by two different HSS (Fig. 5, upper). By combining digestion with different restriction enzymes and different unique K18 probes (Fig. 5, upper), seven different K18 HSS which are correlated with K18 expression were mapped from both directions. Representative Southern blots are shown in Fig. 5, panels A to D. Two weak HSS (a and b) are found associated with the proximal Alu sequence (panel A, lane 4; panel B, lane 1). A strong site or region (c) mapped to the proximal promoter regions (panels A and B, lanes 1 to 4). This is a region of K18 that was expected to be DNase sensitive in permissive tissues because of the location of a typical TATA box element and several potential SP1 binding sites (19). Two weaker HSS (d and f) flanked a stronger site (e) within the first intron which has been shown to contain a complex transcriptional enhancer region (25). One additional and unexpected HSS (g) was found within the sixth exon (panel C, lanes 1 and 4; panel D, lanes 1 to 4 and 7). The location of this HSS is close to the differentially methylated HpaII and HhaI sites within exon 6 (Fig. 1A and B). DNase HSS a and b correlate with transcription of the proximal Alu element. The DNase HSS found distal to the promoter region of K18 map to an Alu repetitive sequence
cis REGULATION OF K18 GENE
VOL. 13, 1993
1819
iv
RNA polymerase III (pol III) promoter elements responsible for the transcription of some Alu sequences (13, 30) sugthat the DNase HSS in this area might reflect an -I -gested 6 shows the additional neighboring transcription unit. Figure tOkb 8kbk 6kb !k to detect designed experiments protection of RNase results II transcripts from the Alu element proximal to the 5' end of K18 in RNAs from tissues of K18 transgenic mice. The 3.7kb Pvu 11 Pvull K18TG1, K18TG2, and K18TG3 transgenic mouse lines 19 of K18. aboutprotected 5, 10, and 18 RNAs copiesfrom integrated liver, Hlncil by the gut andSeveral 5.1kb fragments were Hinc I which express the K18, but not by spleen RNAs in any of the three transgenic lines. As K18 is not expressed in spleen, the ---9 Alu RNAs correlate well with the expression of K18. The major protected fragments (Fig. 6, fragments 2 and 3) correspond closely to the expected size of transcripts initi.4 b ating about 12 bp upstream of the A box of the pol III 0 9 promoter. The same fragments were detected by using a c D B A shorter probe terminating at the XhoI site upstream of the pol III promoter (data not shown), thus indicating the Probe III Probe IV Probe 11 Probe Sp 3_4 6LI7 Kid Kid Sp _Lie_ transcripts protecting the probe were initiated upstream of 2 3 4 5 t the K18 promoter. No specific signal was obtained by using R--23_;; e ~ probes of the opposite orientation. The limitations of map-96-6 ping the a and b sites by Southern blot analysis prevented -4.9 -5.1 -4.4 correlating the b HSS and the pol III promoter elements g -4A4-. more closely. The a site appears to be located within the body of the Alu element. The signal obtained from K18TG2 l-20 C: _ * transgenic gut (Fig. 6, lane 8) was comparable to that of 2 pg -239 d_e, a_-0 of synthetic control RNA (Fig. 6, lane 17). Thus Alu RNA is -1.2 I about 50-fold less abundant than K18 mRNA. However the _-1.2g levels of Alu and K18 RNAs are correlated with regard to both permissive tissues and the number of K18 transgene Pvu51 Hinc 11 Hinc 11 Hinc IS Hinc 11 -4 copies. Furthermore, noAlu RNA was detected in the livers > >< of a related transgenic line (Fig. 6, KpOLac, lanes 14 and 15) FIG. 5. Mapping the position of tirssue-specific DNase I HSS in which contains the same region of K18 but fails to express a K18. The top line shows the position c)f the probes used. The exons LacZ indicator gene in adults (33a). The presence of the a of K18 are represented by filled bowxes, and the two Alu repeat and b HSS correlate with transcription of the Alu element sequences are shown by open boxes. The positions of the DNase I HSS are indicated by vertical arrow: s lettered a to g. Restriction proximal to the K18 promoter and In the direction opposite enzyme fragments of the gene are shovwn as solid lines with the sizes that of K18 transcription. in kilobases above them. The fragmentts generated by particular HSS A regulatory region of K18 is located within exon 6. The g are designated by horizontal arrows a]nd the letter of the site which DNase HSS mapped to exon 6 of K18. This region of the defines one end. Panels A, B, and C represent the same filter gene was tested for regulatory activity by transient expreshybridized sequentially with probes I, I[I, and IV, respectively. Nuclei sion of a series of K18 constructions shown in Fig. 7. All from kidney (Kid), liver (Liv), and spleen (Sp) ofK18 transgenic mice constructions utilized the 5' and 3' flanking sequences, were digested with DNase I or with e ndogenous nucleases (liver) at intron 1, and the polyadenylation signal located in the 3' end 30'C. DNA was purified, digested wit h the restriction enzyme indiof exon 7. Transfection of K18 containing a deletion of cated at the bottom of each panel, arnd subjected to Southern blot analysis using the probes indicated at Ithe top of each panel. Lanes 1 sequences between exons 5 and 7 (construct 1, Fig. 7 and 8A and 2, K18TG2 kidney nuclei digested iwith 5 and 2 g of DNase I per and C) resulted in low levels of K18 RNA. However, ml, respectively, for 5 min. Lanes 3 and 4, K18TG3 liver nuclei substitution of a genomic fragment from the beginning of digested by endogenous nuclease withiout further incubation (lane 3) exon 2 to exon 7 with the corresponding portions of the or for 3 min (lane 4). Lane 5, K18TG2 s pleen nuclei digested with 6 p.g cDNA resulted in effective K18 RNA expression (construct of DNase I per ml for 5 min. Sizes of standards and particular K18 fragments (indented) are shown at the right of each panel. The band 2; Fig. 7 and 8A and C). This suggested that additional regulatory elements were not located in introns 2 to 6. marked with a single asterisk indicatees a DNase-independent fragDeletion of part of the K18 cDNA corresponding to parts of ment unique to K18TG2 mice and likel3y represents either a rearrangeexons 2 and 3 had little effect on expression (construct 3, ment or gene fragment bordering the si te of integration. The fragment Probes:
11
V
a
bc
d
o
._
g
4kb
2kb
Pst
Hinc
III
f
P5t1
1.9kb
11
4.9kb
d
--
Liv
Kid
2
3 4
i
5
2
3
b_a
c -.
_4
t
w
marker with two asterisks is a similax animals. (D) Lanes 1 and 2, K18TG2 liiver nuclei incubated for 5 and 3 min, respectively; lanes 3 and 4, K181FG3 liver nuclei as isolated and with an additional 3-min incubation; ILane 5, K18TG2 spleen nuclei incubated with 6 pLg of DNase I per ml for 5 min. Lanes 6 and 7 contained the same DNA as lanes 3 an d 4 except it was digested with HincII instead of PvuII.
element oriented in the direction opposite that of transcription of K18. The conservation of a Bl-type mouse equivalent of this element in the same orientation and at the same position upstream of mK18 f3-1 aind the conservation of the
Fig. 7 and 8A and C). Deletion of parts of exons 5 and 6 resulted in slightly less efficient expression (construct 4, Fig. 7 and 8A and C). However, two different, nonoverlapping deletions of the cDNA portion containing exon 6 (constructs 5 and 6, Fig. 7 and 8A and C) resulted in about fivefold less K18 RNA after transfection. These results suggest that regulatory activity of this region may be represented by multiple elements. The larger construction that deletes much of exon 6 and most of exon 7 (construct 5, Fig. 7) was rescued by reinsertion of exclusively exon 6 (construct 7, Fig. 8A and C). In addition, the K18 cDNA portion corresponding to only exon 6 was sufficient to ensure a level of
1820
MOL. CELL. BIOL.
NEZNANOV AND OSHIMA
0
Alu RNA Protection K18
B A
Pst
Xho
2 3 45 6 7 A * a*
1
rn
1I
~- ~~I
I---w
Bs/K
TATA n
....
I
AlIu
2 kb
Ex! K1 8 cDNA
Pol 3
247 nt Protected 200 bp
652 nt Probe
op 348 nt Standand
Bs
IIK
BS
__J
__
0-11 TG1
N
mI
P
Gu
1
2
Li Sp G Li 3
4
5
6
TG2
IlI Sp Gu 7
8
TG3
KMLc
Li Sp Gu Li Sp Li Sp t C M 9 10 11 12 13 14 15 16 17 18
1596
-A-JL1.
829 652
p I
d
"
1
2 3 45 6 7
545
1W0
-
*
YJY-~~~~~~~357
472 /
B8S5
41~ ~ ~~~~~~~~~~~~~~~~~1
FIG.~ ~~ ~ ~~
~
w t | ,A,.
~~~~~~~~~~~~~~~~~~~~~~~~j
_
~~~~~~~176
FIG. 6. The Alu sequence proximal to the K18 promoter is transcribed in the direction opposite to that of K18. The top portion shows a map of the region immediately 5' of the first exon (Exl) of K18. The Alu sequence is indicated by the stippled box. Restriction enzyme sites for PstI (Pst) and XhoI (Xho), potential RNA polymerase III promoter elements (A and B), and K18 promoter elements (TATA) are shown. The expected RNA polymerase III start site is indicated (Pol3). The first horizontal arrow represents the 247-nt portion of an RNA polymerase III transcript that would be expected to be protected from digestion by RNase. A single-stranded radioactive RNA probe of 652 nt was hybridized to 10 p.g of total RNA from the indicated source and then subjected to digestion with RNase Tl, electrophoresis under denaturing conditions, and intensifier screen-enhanced autoradiography. A synthetic nonradioactive RNA standard that would yield a 348-nt protected fragment is indicated by the third horizontal arrow (Standard). RNAs were derived from normal, nontransgenic mice (N); K18TG1 (TG1), K18TG2 (TG2), or K18TG3 (TG3) transgenic mice; or a transgenic line which carries a K18-LacZ fusion construction (KpOLac) which includes the portion of K18 indicated but is not expressed in adult animals. P, undigested probe; t, 10 p.g of tRNA; C, 2 pg of synthetic standard RNA and 10 ,ug of tRNA; M, radioactive single-stranded DNA size markers with the sizes indicated on the right side of the figure in nucleotides. The RNAs from gut (Gu), liver (Li), and spleen (Sp) were analyzed. Specific protected fragments are indicated on the left side by the numbers 1 to 4. The size of the most intense protected fragment (fragment 2) was estimated to be 252 nt relative to the DNA standards. Note that the abundance of the protected fragments correlates with K18 RNA expression which is highest in gut, lower in liver, and not found in spleen in the K18 transgenic mice and additionally increases with increasing copy number of K18 (TG3 > TG2 > TG1).
FIG. 7. Structure of K18 portion of plasmids tested by transfection. At the top K18 is represented with exons designated as black boxes and numbered 1 to 7. Open boxes represent the location of Alu sequences. Different constructions are indicated by the circled numbers. Restriction enzyme sites used for deletions of different portions of the gene or cDNA were BstEII (Bs), KpnI (K), PstI (P), BamHI (B), and SmaI (S). Construction 2 substituted the K18 cDNA for the portion of K18 from the beginning of exon 2 to near the end of exon 7. The polyadenylation signal for the K18 RNA is located just downstream of the Kp4nI site and thus was included in all constructions. In constructs 9 and 10, a portion or all of the K18 cDNA representing exon 6 was replaced in frame with the homologous region of K19.
expression nearly equal to that of the whole K18 gene (construct 8, Fig. 7 and 8A and C). Inspection of the sequence of exon 6 reveals several potential regulatory binding motifs (Fig. 9). In an attempt to disrupt the potential regulatory activity of exon 6 while simultaneously minimizing changes to the coding function of the exon, two constructions which substitute portions of K18 exon 6 for the related, homologous coding region of K19 were tested (constructs 9 and 10; Fig. 7 and 8B and C). Both substitutions resulted in decreased transient expression of K18-related RNAs. Construct 10, which substitutes K19 sequences for most of K18 exon 6, was comparable to the most severe deletions. Both K19 substitutions generate open reading frames that retain conserved features of type I keratins. Thus differences between the RNA expression levels of the constructs is not likely due to translational related processes. These experiments localize regulatory activity to exon 6 of K18, a region
cis REGULATION OF K18 GENE
VOL. 13, 1993
A
Construct: P MK 1 2 3 4 5 6 78 t s
1821
C
B K 9 10
652-
544-L
K-P -
472rF
454F 357-
9
z 00
-N *
4.
of 0
Irl
;e
193-
-*-K-K 176-
&
9.
K18 1
2
3
4
5
6
7
8
9
1 0
Construct
129-
FIG. 8. RNA protection analysis of the products of constructions after transient expression in transfected 3T3 cells. DNAs of the indicated constructs (see Fig. 7) were mixed with the pMClneopA standard plasmid and introduced into cells by the calcium phosphate precipitate method. RNA was isolated with care to eliminate all DNA, and 10 mg was analyzed by RNase protection using simultaneous hybridization
with probes for the Neo gene and the K18 first exon. After electrophoresis under denaturing conditions, the protected fragments were detected by intensifier screen-enhanced autoradiographic exposures of X-ray film. (A) DNAs used for transfection are indicated above the lanes. P, probes. The positions of the K18 (K-P) and Neo (N-P) probes are indicated at the left. Size markers are indicated at the far left (in nucleotides). K, K18 gene; t, 10 ,ug of tRNA negative control; s, 10 pg of synthetic Neo and K18 RNAs standards. The positions of the probe fragments protected by Neo mRNA (N) and the K18 mRNAs (K) are indicated at the right. The K18 synthetic standard is slightly shorter than the authentic K18 mRNA. The relative activities of the different constructs are judged by the intensity of the K18 signal relative to the Neo signal. (B) In a separate experiment, constructions 9 and 10 were compared with K18. The bands above the Neo signals represent excess undigested Neo probe. (C) Summary of multiple experiments like those shown in panels A and B. The ratios of the K18 and Neo signals are compared relative to the K18 gene. The error bars indicate the standard deviation of multiple experiments. Values for constructions 9 and 10 represent the means of two experiments, each of which was very similar. Note that constructs 1, 5, 6, 9, and 10 were all much less active than others. All of these delete or alter all or part of exon 6.
which contains a DNase HSS and is differentially methylated. However, insertion of exon 6 sequences either upstream or downstream of either K18 promoter-CAT or TK promoter-CAT constructs failed to enhance transient expression of the CAT reporter gene in 3T3 cells (data not shown). Thus, K18 exon 6 may represent a position-sensitive regulatory region.
DISCUSSION The copy number-dependent and tissue-specific expression of K18 in transgenic mice has permitted us to examine the chromatin state of the gene in different tissues with increased sensitivity and decreased complications from the many K18 pseudogenes found in the human genome (1, 18). The identification of sites or regions of genes which are particularly sensitive to nuclease digestion when packaged as chromatin has been very useful for identifying potential regulatory regions (12). DNase HSS are correlated with non-histone protein-DNA interaction and alterations in nucleosome formation. The seven HSS found in transcriptionally active K18 correlate well with known regulatory regions and identify additional potential regulatory sites. The site located at the proximal promoter was predicted, as tran-
scription complex formation on the TATA box would be expected to exclude nucleosomes and alter chromatin structure. Additionally, HSS within intron 1 may reflect the interaction of multiple proteins with the complex enhancer region (25). Fos and/or Jun proteins are necessary but are not sufficient for the activity of this enhancer region. Additional specific protein-DNA interactions in this area have not yet been identified. The HSS found in exon 6 suggested that this coding region might also have a regulatory function. The independent identification of a HpaII site within exon 6 which is more methylated in spleen than in liver is consistent with a regulatory region which may be recognized by specific DNA-binding proteins. The deletion analysis of this region as tested by transient transfection indicates these sequences can influence expression of K18 dramatically. However, this region, like intron 1, may be the site of binding of more than one transactivating protein because either of two adjacent deletions decreases expression. In addition the activity of this element is not restricted to epithelial cells as the assays were performed in 3T3 fibroblasts that do not express mK18. It is intriguing that an AP1 site (TGAGTCA) which is essential for the enhancer activity of intron 1 is also found within exon 6, 14 bp upstream of the HpaII site which is
1822
NEZNANOV AND OSHIMA
MOL. CELL. BIOL.
T-ag
MstII
K18:AAGGCCAGCTTGGAGAACAGCCTGAGGGAGGTGGAGGCCCGCTACGCCCTACAGATGGAGCAGCTCAACG 10:
GGAAACGGAGGCGCGCTTTGGAGCCCAGCTGGCGCATATCCAGG
BamHI
API
NF1
Smal/HpaII
Hhal
T-ag
K18: GGATCCTGCTGCACCTTGAGTCAGAGCTGGCACAGACCCGGGCAGAGGGACAGCGCCAGGCCCAGGAGTATGAGGCCCTG 10: 9:
CGCTGTACAGCGGTATTGAAGCCCAGCTGGCGGATGTGCGAGCTGATAGTGAGCGGCAGAATCAGGAGTACCAGCGGCTC TACAGCGGTATCGAAGCCCAGCTGGCGGATGTC
K18: CTGAACATCAAGGTCAAGCTGGAGGCTGAGATCGCCACCTACCGCCGCCTGCTGGAAGATGGCGAGGACTTTAA 10:
ATGGACATCAAGTCGCGGCTGGAGCAGGAGATTGCCACCTACCGCAGCCTGCTCGAG
Xhol
EXON 6 AMINO ACID CHANGES
K18:KASLENSLREVEARYALQMEQLNGILLHLESELAQTRAEGQRQAQEYEALLNIKVKLEAEIATYRRLLEDGEDF 10: 9:
TLRETEARFGAQLAHIQALYSGIEAQLADVRADSERQNQEYQRLMDIKSRLEQEIATYRSLLE YSGIEAQLADV
FIG. 9. Sequence of K18 exon 6 (K18 nt 5532 to 5755). Relevant restriction enzyme sites and putative DNA binding protein motifs are indicated and highlighted. T-ag, simian virus 40 T-antigen binding site. AP-1, consensus binding motif for products of the fos and jun gene family of transcription factors. NF1, nuclear factor 1. Note that the deletion of the internal fragment from BamHI to SmaI removes both the AP-1 and NF1 sites. However, deletion of the sequences of exon 6 from the SmaI site to the 3' end of the exon also decreases activity. The exclamation point marks the first nucleotide of the AP1 motif which was changed from the K19 sequence to a C in construction 9.
differentially methylated. However, unlike intron 1, the regulatory activity of exon 6 appears to be position sensitive because insertion of the sequence into CAT vectors driven by the combination of the K18 promoter and intron enhancer failed to influence expression when assayed without supplementary transcription factors (data not shown). Clearly exon 6 was sufficient to rescue a construct containing a deletion of both exon 6 and most of exon 7. Confirmation of the biological importance of this potential regulatory region will require the generation of additional transgenic mice carrying mutations of K18 within this region. The association of DNase HSS with the Alu sequence upstream of the K18 proximal promoter reinforces speculation that conservation of this element upstream of both the human and mouse K18 genes is functionally significant. A similarly oriented Alu sequence participates in the occlusion of an inappropriate promoter upstream of the e-globin gene, thus permitting the appropriate proximal promoter to initiate transcription correctly (36). Similarly, the K18Alu transcription unit may function to insulate the K18 promoter from neighboring transcription units and cis-acting elements. However, it is equally possible that expression of the Alu sequence may only reflect the open chromatin structure dictated by K18 in permissive tissues. While deletion of the proximal Alu element has little effect upon the transcriptional activity of the K18 gene in transient expression tests (25), it remains to be determined whether this putative pol III transcription unit has a function in expression of integrated K18 genes in transgenic mice. DNA methylation was an attractive mechanism to explain the restricted expression of mK18 in cultured differentiated
cell lines. The correlation of transcriptional silence with methylation for both mK18 and mK8 (28, 32) and the observation that treatment of nonexpressing myoblast and fibroblast cell lines with 5-azacytidine could result in the reexpression of both proteins (7, 16) reinforced the view that DNA methylation might be important for regulation of these genes. However, our analysis of both endogenous mK18 and transgenic K18 indicates that differential methylation of the CpG island that includes the first exon does not determine the expression of either gene in liver and spleen. While the differential methylation of regions within intron 1 and exon 6 may be functionally significant, it seems more likely that these differences only reflect neighboring regulatory activity. The tissue-specific DNase HSS found in K18 transgenes must reflect differences in the packaging of the genes in different tissues. Both the pol III promoter elements of the proximal Alu sequence and the TATA box and potential SP1 binding sites would be expected to be recognized by proteins in most if not all tissues. Yet HSS sites corresponding to these areas are absent in spleen and brain nuclei of transgenic mice with more than 10 copies of K18. During development the K18 gene may be designated to be packaged differently in nonexpressing tissue lineages. It will be interesting to determine whether the unusual position-sensitive characteristic of regulatory elements in exon 6 has a role in this process of gene distinction. ACKNOWLEDGMENTS We thank Birgit Lane (Dundee, Scotland) for the K19 cDNA. This work was supported by a grant from the DHHS, NCI, CA42302, and Cancer Center grant CA30199. REFERENCES 1. Abe, M., and R. G. Oshima. 1990. A single human keratin 18 gene is expressed in diverse epithelial cells of transgenic mice. J. Cell Biol. 111:1197-1206. 2. Antequera, F., J. Boyes, and A. Bird. 1990. High levels of de novo methylation and altered chromatin structure at CpG islands in cell lines. Cell 62:503-514. 3. Bird, A. 1986. CpG-rich islands and the function of DNA methylation. Nature (London) 321:209-213. 4. Bird, A., M. Taggart, M. Frommer, 0. J. Miller, and D. Macleod. 1985. A fraction of the mouse genome that is derived from islands of nonmethylated, CpG-rich DNA. Cell 40:91-99. 5. Blumenberg, M. 1988. Concerted gene duplications in the two keratin gene families. J. Mol. Evol. 27:203-211. 6. Brulet, P., C. Babinet, R. Kemler, and F. Jacob. 1980. Monoclonal antibodies against trophectoderm-specific markers during mouse blastocyst formation. Proc. Natl. Acad. Sci. USA 77: 4113-4117. 7. Darmon, M. 1985. Coexpression of specific acid and basic cytokeratins in teratocarcinoma-derived fibroblasts treated with 5-azacytidine. Dev. Biol. 110:47-52. 8. Dente, L., U. Ruther, M. Tripodi, E. F. Wagner, and R. Cortese. 1988. Expression of human a 1-acid glycoprotein genes in cultured cells and in transgenic mice. Genes Dev. 2:259-266. 9. Feinberg, A. P., and B. Vogelstein. 1983. A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem. 132:6-13. 10. Franke, W. W., H. Denk, R. Kalt, and E. Schmid. 1981. Biochemical and immunological identification of cytokeratin proteins present in hepatocytes of mammalian liver tissue. Exp. Cell Res. 131:299-318. 11. Franko, M. C., C. J. Gibbs, D. A. Rhoades, and C. G. Gajdusek. 1987. Monoclonal antibody analysis of keratin expression in the central nervous system. Proc. Natl. Acad. Sci. USA 84:34823485. 12. Gross, D. S., and W. T. Garrard. 1988. Nuclease hypersensitive sites in chromatin. Annu. Rev. Biochem. 57:159-197.
VOL. 13, 1993 13. Howard, B. H., and K. Sakamoto. 1990. ALU interspersed repeats: selfish DNA or a functional gene family? New Biol. 2:759-770. 14. Jackson, B. W., C. Grund, E. Schmid, K. Burke, W. Franke, and K. Ilimensee. 1980. Formation of cytoskeletal elements during mouse embryogenesis. Intermediate filaments of the cytokeratin type and desmosomes in preimplantation embryos. Differentiation 17:161-179. 15. Julien, J.-P., I. Tretjakoff, L. Beaudet, and A. Peterson. 1987. Expression and assembly of a human neurofilament protein in transgenic mice provide a novel neuronal making system. Genes Dev. 1:1085-1095. 16. Knapp, A. C., F. X. Bosch, M. Hergt, C. Kuhn, S. WinterSimanowski, E. Schmid, S. Regauer, J. Bartek, and W. W. Franke. 1989. Cytokeratins and cytokeratin filaments in subpopulations of cultured human and rodent cells of nonepithelial origin: modes and patterns of formation. Differentiation 42:81102. 17. Kulesh, D. A., G. Cecena, Y. M. Darmon, M. Vasseur, and R. G. Oshima. 1989. Posttranslational regulation of keratins: degradation of mouse and human keratins 18 and 8. Mol. Cell. Biol. 9:1553-1565. 18. Kulesh, D. A., and R. G. Oshima. 1988. Cloning of the human keratin 18 gene and its expression in nonepithelial mouse cells. Mol. Cell. Biol. 8:1540-1550. 19. Kulesh, D. A., and R. G. Oshima. 1989. Complete structure of the gene for human keratin 18. Genomics 4:339-347. 20. Melton, D. A., P. A. Krieg, M. R. Rebagliati, T. Maniatis, K. Zinn, and M. R. Green. 1984. Efficient in vitro synthesis of biologically active RNA and RNA hybridization probes from plasmids containing a bacteriophage SP6 promoter. Nucleic Acids Res. 12:7035-7055. 21. Moll, R., W. W. Franke, D. L. Schiller, B. Geiger, and R. Krepler. 1982. The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors and cultured cells. Cell 31:11-24. 21a.Neznanov, N., and R. G. Oshima. Unpublished results. 22. Oshima, R. G. 1981. Identification and immunoprecipitation of cytoskeletal proteins from murine extra-embryonic endodermal cells. J. Biol. Chem. 256:8124-8133. 23. Oshima, R. G. 1982. Developmental expression of murine extra-embryonic endodermal cytoskeletal proteins. J. Biol. Chem. 257:3414-3421. 24. Oshima, R. G. 1992. Intermediate filament molecular biology. Curr. Opin. Cell Biol. 4:110-166. 25. Oshima, R G., L. Abrams, and D. Kulesh. 1990. Activation of an intron enhancer within the keratin 18 gene by expression of
cis REGULATION OF K18 GENE
1823
c-fos and c-jun in undifferentiated F9 embryonal carcinoma cells. Genes Dev. 4:835-848. 26. Oshima, R. G., W. E. Howe, F. G. Klier, E. D. Adamson, and L. H. Shevinsky. 1983. Intermediate filament protein synthesis in preimplantation murine embryos. Dev. Biol. 99:447-455. 27. Oshima, R. G., J. L. Millan, and G. Cecena. 1986. Comparison of mouse and human keratin 18: a component of intermediate filaments expressed prior to implantation. Differentiation 33:6168. 28. Oshima, R. G., K. Trevor, L. H. Shevinsky, 0. A. Ryder, and G. Cecena. 1988. Identification of the gene coding for the Endo B murine cytokeratin and its methylated, stable inactive state in mouse nonepithelial cells. Genes Dev. 2:505-516. 29. Parfett, M., R. Hofbaner, K. Brudzynski, D. R. Edwards, and D. T. Denhardt. 1989. Differential screening of the cDNA library with cDNAs probes amplified in a heterologous host: isolation of murine grp78 (BiP) and other serum regulated low abundance mRNAs. Gene 82:291-303. 30. Perez-Stable, C., and C.-K. J. Shen. 1986. Competitive and cooperative functioning of the anterior and posterior promoter elements of an Alu family repeat. Mol. Cell. Biol. 6:2041-2052. 31. Sambrook, J., E. F. Fritsch, and T. Maniantis. 1989. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 32. Tamai, Y., Y. Takemoto, M. Matsumoto, T. Morita, A. Matsushiro, and M. Nozaki. 1991. Sequence of the EndoA gene encoding mouse cytokeratin and its methylation state in the CpG-rich region. Gene 104:169-176. 33. Thomas, K. R., and M. R. Capecchi. 1987. Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells. Cell 51:503-512. 33a.Thorey, I., J. J. Meneses, N. Neznanov, R. A. Pedersen, and R. G. Oshima. Unpublished results. 34. Vassar, R., M. Rosenberg, S. Ross, A. Tyner, and E. Fuchs. 1989. Tissue-specific and differentiation-specific expression of a human K14 keratin gene in transgenic mice. Proc. Natl. Acad. Sci. USA 86:1563-1567. 35. Wu, C. 1989. Analysis of hypersensitive sites in chromatin. Methods Enzymol. 170:281-284. 36. Wu, J., G. J. Grindlay, P. Bushel, L. Mendelsohn, and M. Allan. 1990. Negative regulation of the human a-globin gene by transcriptional interference: role of an Alu repetitive element. Mol. Cell. Biol. 10:1209-1216. 37. Zimmerman, K., E. Legouy, V. Stewart, R. DePinho, and F. W. Alt. 1990. Differential regulation of the N-myc gene in transfected cells and transgenic mice. Mol. Cell. Biol. 10:2096-2103.
