Low Low High Medium Low Low Low Low High High High High High High High High High High High High High High High Low Low Low ND Low LOW

We have introduced sequences encoding the lac repressor of Escherichia coli into the genome of the mouse. One sequence was derived from the bacterial lac operon and the other was created by reencoding the amino acid sequence of l a d with mammalian codons. Both versions are driven by an identical promoter fragment derived from the human 0-actin locus and were microinjected into genetically identical pronuclear stage embryos. All transgenes utilizing the bacterial coding sequence were transcriptionally silent in all somatic tissues tested. The sequence reencoded with mammalian codons was transcriptionally active at all transgene loci and expressed ubiquitously. Using methylation-sensitive enzymes, we have determined the methylation status of lac repressor transgenes encoded by either the bacterial or mammalian sequence. The highly divergent bacterial sequence was hypermethylated at all transgene loci, while the mammalian sequence was only hypermethylated at a high copy number locus. This may reflect a normal process that protects the genome from acquiring new material that has an abnormally divergent sequence or structure. T HE introduction of foreign genes and DNA fragments into the genome of the mouse has led to the elucidation of the function of many normal genes and to an understanding of how mutations in particular genes disrupt phenotype. The ability to introduce exogenous DNA sequences that code for either normal or mutant gene products, however, has been limited to those whose expression results in benign or, at the least, sublethal phenotypic changes. In an effort to circumvent problems that arise from the unregulated expression of introduced sequences, we have been constructing a regulatable transgenic system for the mouse that is based on the lactose (lac) operon of Escherichia coli (JACOB and MONOD 1961). Like analogous systems that use temperature-sensitive mutations to study lethal mutations in bacteria and lower eukaryotes, this system would allow the introduction and analysis of embryonic lethal genes at the organismal level without compromising the viability of their host. The capacity to regulate genes in the mouse in vivo would greatly expand the repertory of genes that can be altered and analyzed within the context of an organism more closely related to the human. Use of regulatory elements derived from the bacterial lac operon in our transgenic system required that we introduce sequences coding for the lac repressor (lacl) that would be expressed in the mouse. Data from many other experiments in which bacterial genes have been introduced into the mammalian genome have implicated methylation in the inability of these genes to be Curresponding author: Heidi Scrable, Department of Neuroscience, University of Virginia, Charlottesville, Virginia 22908. E-mail: [email protected] Genetics 147: 297-304 (September, 1997) expressed. Sporadic expression of reporter genes like lacZin cultured mouse embryos (NILSSON and LENDAHL 1993) and in the transgenic mouse (BEDDINGTON et al. 1989; reviewed in CUI et al. 1994) has been attributed to the gradual acquisition of methylated sequences at transgenic as well as endogenous loci as the methylation patterns characteristic of the mature genome are stablished and maintained. To avoid suppression of gene activity by methylation directed at the bacterial l a d coding sequence, we reencoded the amino acid sequence of the lac repressor with mammalian codons (ZHANG et al. 1991). To test the effect on expression of altering only the coding sequence, we prepared two transgene constructs that were identical except for the sequence encoding l a d and microinjected them into genetically identical mice. A comparison of gene activity and methylation status of the two types of repressor transgene suggests that during the process of transgenesis the host genome can respond differentially to introduced sequences. This response appears to depend on how closely the introduced sequences resemble the host genome and results in transcriptional activation of the closely related mammalian l a d sequence and transcriptional suppression of the widely divergent bacterial sequence. MATERIALS AND METHODS Synthesis of mammalian lacl: Segments of 80 nucleotides in length corresponding to either strand of the gene were synthesized on an Applied BioSystems DNA synthesizer, annealed, and ligated in groups of three overlapping fragments. These larger fragments ["240 base pairs (bp)] were ligated to each other in turn to produce the final product. The mam298 H. Scrable and P. J. Stambrook atg aaa cca gta aca gcc ggt gtc tct tat aac cag gcc agc cat ccc aac aga gtg gca aaa gtg gaa gca gcc agc ttg ctg att gga ctg cat gca cca tct tct aga gct gat caa tca atg gta gaa aga gct gca gtg cac aat ggg ctg atc att aac gcc att gct gtg gaa gca ctc ttt ctt gat ctg ggt gtg gag cat aac agt att att ttc caa att gca ctg ctt tca gca agg ctg aga ctc act agg aat caa gaa g c tug aut acc atg caa atg ctg gca atg aga gcc att gca atg ctg gtt gcc ggt gca gat atc tca gaa gac agc tca tgt atc aaa caa gat ttt gtg gac aga ttg ctg gtg aag ggc aat cag gcc tct ccc cgg gca aag aga aaa acc acc Eta gca aga cag gtt gca gct ctg ccc aag A A L P K gtt gca tcc aga aaa acc ctg aat gca ggc tcc agt gca gcc tct gtg gaa gcc caa aga gat gac act aat cag aca gat ggt ttg gga ctc agt tgg cat ata gct ggg ttt att gtt atg gca uucl ctu tat gat ccc tca ugg caa caa ggc gtc tca aat aca tca ctc gaa agt aag gtg K V FIGURE 1.-DNA sequence of mammalian lacl. Each triplet encodes the same amino acid found in the bacterial lac repressor. The translation product of the nuclear localization signal from SV40 is indicated beneath the 30-bp extension appended to the 3' end of the lacl coding region immediately 5' to STOP (tga). malian Zuclsequence was also modified to include the nuclear localization signal (NLS) from SV40, which results in the sequestration of most repressor protein in the nucleus, as described in LIU et al. (1992). The NLS was appended to the carboxy terminus of the protein, with two additional alanines serving as a linker. The sequence was also designed to avoid inclusion of potential splice donor or splice acceptor sites. The final product was sequenced and several cloning artifacts corrected by PCR-based, sitedirected mutagenesis (BOWMAN et al. 1990). The sequence of the mammalian lac repressor is given in Figure 1 above, with the translation product of the NLS and alanine linker indicated at the 3' end. Production of transgenic mice: Mice transgenic for the bacterial gene coding for the lac repressor (lucl) were made by microinjection into (C57BL/6 X SJL)F, hybrid pronulear stage embryos. Thirteen founders and their offspring were analyzed for the data used in this article (01-14, Table 1). In addition, to circumvent anticipated problems with expression of lacl as a result of methylation, we attempted to derive lad transgenics in inbred DBA/2J embryos. DBA/2J mice have been reported to hypomethylate transgene DNA (ENGLER et al. 1991). In a typical attempt, 325 eggs were harvested from 30 superovulated DBA/2J females; 100 of the 325 eggs had been fertilized, and 0 of 100 survived implantation and/or embryogenesis. We switched to hybrids of DBA/2J with SJL, another hypomethylating strain (ENGLER et al. 1991). Fortyseven eggs from DBA/2J females fertilized in vitro with sperm from SJL males were microinjected and cultured overnight. Thirty-four progressed to the two-cell stage and were implanted into isogenic foster mothers. Three animals were born and two of them were transgenic. One founder and the line derived from it were analyzed for the data used in this article (D2, Table 1). DNA extraction: DNA was extracted from tissue by lysis in 10 mM Tris, pH 8; 10 mM NaCl; and 1 mM EDTA in the presence of 1% sarkosyl and proteinase K (100-200 pg/ml). Proteins and nucleic acids were separated by extracting three times with phenol and once with phenol/CHCl,/isoamyl alcohol (25:24:1). Nucleic acids were precipitated from the aqueous phase by adding 0.5 volumes of 7.5 M ammonium acetate, and then 1 volume isopropanol. High molecular weight DNA was spooled onto a glass rod and transferred to TE (10 mM Tris/l mM EDTA) and resuspended overnight. The DNA was reprecipitated the next day by the addition of 0.1 volume 3 M sodium acetate and 2 volumes of ethanol precooled to -Zoo, washed in 70% ethanol, and resuspended in TE. DNA was extracted from mouse tail by digesting 1 cm of tissue at 58" overnight in 700 p150 mM Tris-HC1, pH 8/100 mM EDTA, pH 8/100 mM NaC1/1% SDS to which was added 35 p1 10 mg/ml Proteinase K (BRL, Bethesda, MD) just prior to incubation. Following digestion with 20 pl of RNAseA (Sigma, St. Louis, MO; 10 mg/ml made up in water, then boiled 15 min to destroy DNAse) for 2 hr at 37", the samples were extracted with equal volumes of phenol (3X), phenol/CHCll/isoamyl alcohol (25:24:1) ( Z X ) , or CHCIS/isoamyl alcohol (24:l) (1 X). The DNA was precipitated out of the aqueous phase by adding 0.5 M NH,OAc to the top of the tube, spooled onto a glass rod, and suspended in 500 pl TE-4 (10 mM Tris/O.l mM EDTA). Southern blotting: DNA (5 pg) was electrophoresed through 0.8% agarose after digestion with appropriate restriction endonucleases (GIBCO/BRL). Transfer was to HybondN+ (Amersham) in 0.25 M NaOH/1.5 M NaCl following treatment of the gel in 0.25 M HCl (2 X 15 min) and 0.5 M NaOH/ 1.5 M NaCl (30 min). The resultant Southern blot was dried at 80" for 10 min and then was prehybridized overnight at 42" in 5X SSC, 1OX Denhardt's, 0.05 M phosphate (pH 6.7), 1% SDS, 5% dextran sulfate, and 50% formamide, with 500 pg ml" boiled salmon sperm DNA added. The blots were hybridized either at 42" or at 47" in 5X SSC, 2X Denhardt's, 0.02 M phosphate (pH 6.7), 1% SDS, 10% dextran sulfate, and 50% formamide, with 200 pg/ml salmon sperm DNA and 2 X lo6 cpm/ml labeled probe added. Probes were labeled by random priming and were usually isolated inserts. After 36 hr hybridization, the blots were washed in 0.1 X SSC/O.l% SDS, either at 55" or at 65". The blots were subjected to autoradiography with Dupont/NEN Reflection film and intensifying screens. RNA extraction and Northern blotting: Tissue was pulverized in a liquid nitrogen-cooled mortar and total RNA was extracted by the acid phenol method using TRI Reagent (Molecular Research Products, Inc., Cincinnati, OH). RNA was electrophoresed through a 1% agarose/0.67 M formaldehyde gel after formaldehyde denaturation for 15 min at 55". RNA was transferred to a nylon membrane (Hybond-N; Amersham) by capillary blotting in 2OX SSC and fixed to the membrane by UV crosslinking (Stratalinker; Stratagene). Blots were prehybridized overnight at 42" in 50% formamide, 5X Denhardt's, 5X SSPE, and 0.1% SDS, with 200 pg/ml boiled salmon sperm DNA and 1 ,ug/ml poly(A) added. Blots were hybridized at 42" in 50% formamide, 1X Denhardt's, 0.04 M phosphate (pH 6.7), and 0.05% SDS, with DNA and poly(A) added, as above, and 2 X loG cpm/ml a["P]dCTP labeled probes. Probes: Probes were labeled by the method of random priming (FEINBERG and VOGELSTEIN 1983). lad probes were derived from either the bacterial sequence contained in pCMvlacI or the mammalian sequence contained in the plasmid pI5, both in the plasmid collection of the Stambrook lab. The p-actin probe was derived from a murine cDNA (BUCKINGHAM 1985) given to us by MARGARET BUCWNGHAM. The MyoD probe was a murine cDNA in plasmid p VZC13p (DAVIS et al. 1987) given to us by ANDREW LASSAR, and p53 w d S a gift from ALAN BERNSTEIN (ROVINSKI ~t al. 1987). lac Repressor in Transgenic Mice