Development of an inducible expression system in mice. (a) A schematic representation of our strategy to generate the R26STOPrtTA allele. The targeted locus contains a pgk-neo gene (neo) for positive selection and a triple polyadenylation site (a STOP sequence indicated by blue arrowhead), both flanked by loxP sites, preceding the rtTA2^S-M2 gene, which is followed by a triple polyadenylation sequences. The STOP sequence is removed by Cre-mediated recombination between loxP sites (red triangles) to activate the expression of rtTA. Diphtheria toxin (DTA) also was used for negative selection (16). V, EcoRV; S, SacII; X, XbaI. (b) Southern and PCR analysis. Using a 5′-external probe, the EcoRV-digested wild-type band (wt, 11.0 kb) or a smaller-sized mutant band (m, 3.8 kb) was detected by Southern blot analysis. Mice heterozygous and homozygous for the R26STOPrtTA allele were analyzed by PCR for the 5′ and 3′ junction of the wt and mutant alleles, as indicated. The presence of neo in the R26STOPrtTA mice also was confirmed by PCR analysis. (c-f) Expression of the lacZ reporter in triple transgenic embryos (β-actin-Cre, R26STOPrtTA and TRE-lacZ) at E10.5 (c) or E12.5 (d) was analyzed by whole-mount β-gal staining at 37°C for 16 h. (Right) Dox induction for 3 days or (Left) without Dox induction. (e and f) Sections of the stained embryos shown in c, counter-stained with nuclear fast red, revealed a uniform expression of the reporter in the presence (f) but not in the absence (e) of Dox.

Lineage-specific gene expression in CNC derivatives in vivo and in vitro. (a) Our strategy for conditional expression of a target gene using the dual-level expression system. Tissue or lineage specificity is provided by Cre-mediated excision of the loxP-flanked STOP sequence, and further temporal activation is controlled by Dox administration. The expression of rtTA in this system becomes independent of regulation by the tissue-specific promoter (TSP) after the excision of the loxP-flanked STOP sequence. The R26STOPrtTA gene activated by Cre remains active in all of the daughter cells even after Cre expression has been turned off. (b) E11.5 Wnt1-Cre/R26STOPrtTA/TRE-lacZ triple transgenic embryos, induced with Dox for 5 days (embryo, Upper Left) or without Dox induction (other three embryos) (β-gal staining for 2 h). (c) Enlargement of Dox-induced embryo from a, showing expression in derivatives of the Wnt1-expressing neural crest cells, including maxilla, mandible, cranial nerves, skull, eyes, skin, and salivary glands. (d-i) Sections of the β-gal-stained E11.5 embryo showing targeted gene expression in neural crest derivatives, including the lateral (LNP) and medial (MNP) nasal processes (d), the mandibular (Mn) and maxillary (Mx) components of branchial arch (e and f), the optic cup (f, Oc), the trigeminal (g, TG) and dorsal root ganglia (h, DRG), the dorsal neural tube (h, NT), the united ganglion of XI and X nerves (black arrowhead), the inferior ganglion of XI nerve (black arrow), and the branches of VII nerve (white arrowhead) (i). β-Gal staining of a triple transgenic E16.5 embryo (k and l) and 21-day-old mouse (n) showed that conditional gene expression could be manipulated in the CNC-derived skeletal elements of the developing mouse skull vault. Control mice without Dox induction revealed no expression of lacZ (j and m). f, frontal bone; ip, interparietal bone; n, nasal bone; nc, nasal cartilage; p, parietal bone. Sections of the β-gal-stained E16.5 embryos revealed that the lacZ reporter is expressed in the cranial sutures (o, arrow) and salivary glands (p). The neural crest-derived osteoblasts isolated from nasal and frontal bones were cultured in medium containing (r) or lacking (q) Dox for 16 h and analyzed by β-gal staining (at 37°C for 1 h) for inducible expression of the target gene. Absolutely no activation of the TRE-lacZ reporter (0%) was observed in cells without Dox. In contrast, 95% of the neural crest-derived osteoblasts exhibited nuclear β-gal staining upon Dox treatment. A similar expression pattern was observed in calvarial osteoblasts isolated from the newborn carrying Wnt1-Cre/R26RlacZ transgenes. No lacZ expression was detected in osteoblasts isolated from the parietal bone (s), whereas ≈95% of the cells were positive for β-gal staining (at 37°C for 2 h) in the nasal/frontal bone-derived osteoblasts (t). [Scale bars: 3 mm (b, m, and n); 1 mm (c); 100 μm (d-g and i); 200 μm (h, and o-t); and 2 mm (j-l).]

Conditional gene expression in a spatial- and temporal-specific fashion. (a) Nestin-Cre/R26STOPrtTA/TRE-lacZ triple transgenic E13.5 embryo (Right) showing expression of the lacZ reporter in the neuroepithelial lineage and control double transgenic (R26STOPrtTA and TRE-lacZ) E13.5 embryos (Left). Both embryos were analyzed by β-gal staining at 37°C for 2 h, after 5 days of Dox induction. (b) Section of the β-gal-stained triple transgenic embryo from a showing targeted gene expression in the dorsal neural epithelium. A similar pattern of expression, in only the dorsal neural tube, was also observed in Nestin-Cre/R26RlacZ embryos (data not shown). (c and d) Whole-mount GFP analysis (exposure time, 2 sec) of thymus of a triple transgenic mouse carrying Lck-Cre/R26STOPrtTA/TRE-Axin-GFP revealed that the target gene was conditionally activated upon Dox treatment for 7 days (d, right), but not the control without Dox (d, left). A phase-contrast image of the thymuses is shown in c. (e-h) FACS staining analysis, further demonstrating the expression of GFP (g and h) in the CD4 and CD8 double-positive T cells (e and f) isolated from the Lck-Cre/R26STOPrtTA/TRE-Axin-GFP thymus after Dox induction for 3 days (f and h) but not the control thymus (e and g). M1, 0.35% (e); 90.11% (f). [Scale bars, 2 mm (a), 200 μm (b), and 500 μm (c and d).]

Manipulating gene expression in the mammary gland. (a) A schematic diagram illustrates the strategy for targeted expression. The WAP-Cre transgene, which is active in late pregnancy, turns on the expression of rtTA in differentiating epithelia and differentiated alveoli. The rtTA remains expressed in derivatives of the WAP-expressing cells (MEP progenitors and their progeny), even though Cre is no longer present in the involved gland. Diagram modified from Wagner et al. (22). β-Gal staining (37°C for 5 h) of the WAP-Cre/R26STOPrtTA/TRE-lacZ parous gland revealed no expression of the target gene without Dox (b) and induction of expression with Dox for 7 days (c). The TRE-lacZ transgene was conditionally activated in the recently identified MEP progenitors of the triple transgenic female 2 weeks after weaning/mammary gland remodeling. A section of the β-gal-stained parous gland was counterstained with nuclear fast red (d). β-Gal staining (37°C for 1 h) of the primary mammary epithelia, which were isolated from the WAP-Cre/R26STOPrtTA/TRE-lacZ parous gland 4 weeks after weaning, revealed nuclear staining of the MEP cells (e). The cultured cells were subsequently stained by immunostaining with anti-keratin8/18 antibody (brown) to confirm identify of the MEP cells. This system enables manipulation of gene expression in a subpopulation of the MEP cells in vivo and in vitro. [Scale bars: 1 mm (b and c) and 50 μm(d and e).]