We are indebted to the Functional Genomics Center Zurich (Zurich, CH) for financial as well as technical support (thanks especially to Andrea Patrignani, Dr Ulrich Wagner, and Dr Hubert Rehrauer for their assistance in data generation and evaluation), and to Dr Michael Leviten (San Carlos, CA) for the help in generating Mtf1 conditional knockout mice. We also thank Prof. Ueli Schibler (Geneva, CH) for his advice on the ADDER technique, Prof. Michel Aguet (Epalinges-Lausanne, CH) for the gift of Mx-Cre mice, Dr George Hausmann and Dr Michael Fetchko for critical reading of the manuscript, and Fritz Ochsenbein for preparing the figures. Funding to pay the Open Access publication charges for this article was provided by the Kanton Zurich.
Conflict of interest statement. None declared.

Figures and Tables
Figure 1  Deletion of Mtf1 in adult mouse liver. (a) Generation of Mtf1 conditional knockout mice. The targeted allele was obtained by homologous recombination of wild-type (wt) allele and targeting vector in ES cells. Removal of the neomycin cassette (NEO) by Cre recombinase led to the conditional knockout allele Mtf1loxP. Conditional Cre-mediated deletion of exons 3 and 4 (Mtf1Δ) results in loss of function via loss of an essential part of the DNA-binding domain and the generation of a new stop codon right after exon 2. Exons 3 to 7 of Mtf1 are indicated by grey boxes, loxP sites by black triangles. TK, thymidine kinase cassette. Restriction enzymes: San, SanDI; B, BbvCI; S, SrfI; H, HpaI. The HpaI site indicated by the crossed H was lost during the cloning procedure for the targeting vector. (b) RT–PCRs with total liver RNA from pI–pC-induced male Mtf1Mx-cre or Mtf1loxP mice. The used primer pair results in products of 589 bp and 218 bp with full-length mRNA and mRNA without exons 3 and 4, respectively. (c) EMSA with liver protein extract of a pI–pC-induced male Mtf1Mx-cre or Mtf1loxP mouse. MTF-1 protein–DNA complex formation was tested with 32P-labeled MRE consensus oligonucleotide MRE-s. Specificity of binding was verified with excess of unlabeled competitor MREd or unrelated Gal4 oligonucleotide; Sp1 bandshifts with 32P-labeled Sp1 consensus oligonucleotide were included as a loading control.
Figure 2  Sepw1 basal expression depends on MTF-1. (a) Semiquantitative RT–PCRs for Sepw1 mRNA using total liver RNA from pI–pC-induced male Mtf1Mx-cre or Mtf1loxP mice. The animals had obtained either mock s.c. injections (−Cd) or s.c. injections with 20 µmol/kg body weight CdSO4 (+Cd) 6 h before sacrificing them. RT–PCRs for Hprt mRNA were used as internal control to adjust the amount of total RNA used. (b) S1 analysis for Sepw1 mRNA with RNA described in (a), using a 32P-labeled Sepw1 S1 probe. A 32P-labeled S1 probe for Hprt mRNA was used to adjust the amount of RNA used. (c) MRE core consensus sequences TGCRCNC (bold letters) and flanking sequences found in a region of 1000 bp upstream from Sepw1 transcription start; the position of each core sequence is indicated. (d) EMSA with liver protein extracts of a male Mtf1loxP or a pI–pC-induced, male Mtf1Mx-cre mouse, both mock-treated. MTF-1 protein–DNA complex formation was tested with 32P-labeled Sepw1 MRE1 or MRE2 oligonucleotide, respectively. Specificity of binding was verified with excess of unlabeled competitor MREd or unrelated Gal4 oligonucleotide. 32P-labeled MRE-s was included to indicate the position of an MTF-1-DNA complex; bandshifts for the common transcription factor Sp1 with 32P-labeled Sp1 consensus oligonucleotide were obtained as protein loading control.
Figure 3  Cadmium response of Ndrg1 depends on MTF-1. (a) Semiquantitative RT–PCRs for Ndrg1 mRNA using total liver RNA from pI–pC-induced male Mtf1Mx-cre or Mtf1loxP mice. The animals had obtained either mock s.c. injections (−Cd) or s.c. injections with 20 µmol/kg body weight CdSO4 (+Cd) 6 h before sacrificing them. RT–PCRs for Hprt mRNA were used as internal control to adjust the amount of total RNA used. (b) MRE core consensus sequences TGCRCNC (bold letters) and flanking sequences found in a region of 1000 bp upstream from Ndrg1 transcription start; the position of each core sequence is indicated. (c) EMSA with liver protein extracts of a male Mtf1loxP or a pI–pC-induced, male Mtf1Mx-cre mouse, both mock-treated. MTF-1 protein–DNA complex formation was tested with 32P-labeled Ndrg1 MRE1 or MRE2 oligonucleotide, or a 32P-labeled oligonucleotide including both MRE3 and MRE4 (MRE3,4). Specificity of binding was verified with excess of unlabeled competitor MREd or unrelated Gal4 oligonucleotide. 32P-labeled MRE-s was included to indicate the position of an MTF-1-DNA complex; Sp1 bandshifts with 32P-labeled Sp1 consensus oligonucleotide were obtained as protein loading control.
Figure 4  Cadmium response of Csrp1 depends on MTF-1. (a) Semiquantitative RT–PCRs for Csrp1 mRNA using total liver RNA from pI–pC-induced male Mtf1Mx-cre or Mtf1loxP mice. The animals had obtained either mock s.c. injections (−Cd) or s.c. injections with 20 µmol/kg body weight CdSO4 (+Cd) 6 h before sacrificing them. RT–PCRs for Hprt mRNA were used as internal control to adjust the amount of total RNA used. (b) MRE core consensus sequences TGCRCNC (bold letters) and flanking sequences found in a region of 1000 bp upstream from Csrp1 transcription start; the position of each core sequence is indicated. (c) EMSA with liver protein extracts of a male Mtf1loxP or a pI–pC-induced, male Mtf1Mx-cre mouse, both mock-treated. MTF-1 protein–DNA complex formation was tested with 32P-labeled Csrp1 MRE1, MRE2, MRE3 or MRE4 oligonucleotide, respectively. Specificity of binding was verified with excess of unlabeled competitor MREd or unrelated Gal4 oligonucleotide. 32P-labeled MRE-s was included to indicate the position of an MTF-1–DNA complex; Sp1 bandshifts with 32P-labeled Sp1 consensus oligonucleotide were obtained as protein loading control.
Figure 5  MTF-1 represses basal expression of Slc39a10. (a) Semiquantitative RT–PCRs for Slc39a10 mRNA using total liver RNA from pI–pC-induced male Mtf1Mx-cre or Mtf1loxP mice. The animals had obtained either mock s.c. injections (−Cd) or s.c. injections with 20 µmol/kg body weight CdSO4 (+Cd) 6 h before sacrificing them. RT–PCRs for Hprt mRNA were used as internal control to adjust the amount of total RNA used. (b) MRE core consensus sequences TGCRCNC (bold letters) and flanking sequences found in a region of 1000 bp upstream from Slc39a10 transcription start; the position of each core sequence is indicated. (c) EMSA with liver protein extracts of a male Mtf1loxP or a pI–pC-induced, male Mtf1Mx-cre mouse, both mock-treated. MTF-1 protein–DNA complex formation was tested with 32P-labeled Slc39a10 MRE1 or MRE2 oligonucleotide, respectively. Specificity of binding was verified with excess of unlabeled competitor MREd or unrelated Gal4 oligonucleotide. 32P-labeled MRE-s was included to indicate the position of an MTF-1–DNA complex; Sp1 bandshifts with 32P-labeled Sp1 consensus oligonucleotide were obtained as protein loading control.
Figure 6  Cells with reduced glutathione level that also lack MTF-1 are hypersensitive to cadmium. The viability of cells was assessed with the so-called MTT assay. Mouse embryonic fibroblasts with (ckoC) and without (delC19, delC21 and delC23) functional Mtf1 were compared. Cells were pre-incubated in medium containing 0, 5, 10, 25 or 50 µM BSO for 24 h and further exposed to 0, 5, 10 or 20 µM CdCl2 (Cd) in the specified pre-incubation medium for an additional 24 h. Results are expressed as mean values ± SD (n = 3) normalized to the respective value of untreated cells.
Table 1  Comparison of liver gene expression for pI–pC-induced Mtf1Mx-cre and Mtf1loxP mice (up- or downregulation at least 2-fold, P ≤ 0.05)
The animals had obtained either mock s.c. injections (−Cd) or s.c. injections with 20 µmol/kg body weight CdSO4 (+Cd) 6 h before sacrificing them. The expression values for each gene are given as mean value of three animals per group, normalized to the mean value of group Mtf1loxP −Cd (relative activity). Grey shading indicates the two groups of animals compared, relative activities for the other two groups are shown for a complete overview. The number of MRE core consensus sequences TGCRCNC in a region of 1000 bp upstream from the annotated transcription start is indicated.
aOnly incomplete region up to −1000 bp from transcription start is available in database.
bMean values of two independent Affymetrix probe sets.
cMean value of four independent Affymetrix probe sets.
Table 2  Comparison of liver gene expression for cadmium- and mock-treated mice (up- or downregulation at least 2-fold, P ≤ 0.05)
Gene symbol  Gene title  Relative activity  P-value
Mtf1loxP  Mtf1loxP  Mtf1Mx-cre  Mtf1Mx-cre
−Cd  +Cd  −Cd  +Cd
(a) Genes upregulated in cadmium-treated mice
    Cbr3  Carbonyl reductase 3  1  19.600  1.430  27.650  0.002
    Npn3a  Neoplastic progression 3  1  12.907  1.045  12.460  0.002
    Ddc  Dopa decarboxylase  1  7.166  1.776  7.633  0.024
    Serpina9  Serine (or cysteine) proteinase inhibitor, clade A, member 9  1  6.345  0.998  9.455  0.041
    Ppfibp2  Protein tyrosine phosphatase, receptor-type, F interacting, binding protein 2  1  3.771  1.167  5.279  0.004
    Gclc  Glutamate-cysteine ligase, catalytic subunit  1  3.678  1.166  4.046  0.002
    Pgda  Phosphogluconate dehydrogenase  1  3.530  1.419  3.821  0.022
    Ikbkgb  Inhibitor of kappaB kinase gamma  1  3.319  1.128  3.702  0.010
    Txnrd1  Thioredoxin reductase 1  1  3.273  1.046  3.467  0.005
    Kdelr2  KDEL (Lys-Asp-Glu-Leu) endoplasmic reticulum protein retention receptor 2  1  3.209  1.138  3.229  0.006
    Ddit41  DNA-damage-inducible transcript 4-like  1  2.906  1.083  4.713  0.049
    Aqp8  Aquaporin 8  1  2.810  1.505  3.528  0.049
    Gstm4  Glutathione-S-transferase, mu 4  1  2.614  1.346  2.800  0.037
    Bag3  Bcl2-associated athanogene 3  1  2.528  0.997  3.905  0.049
    Gsrc  Glutathione reductase 1  1  2.376  1.224  2.611  0.010
    Pir  Pirin  1  2.235  1.011  2.595  0.043
    Htatip2  HIV-1 tat interactive protein 2, homolog (human)  1  2.130  1.069  2.585  0.035
    Mocos  Molybdenium cofactor sulfurase  1  2.114  1.201  2.424  0.026
    Abcc4  ATP-binding cassette, subfamily C, member 4  1  2.095  0.973  3.072  0.030
    Rassf6  Ras associated (RalGDS/AF-6) domain family 6  1  2.056  1.087  2.930  0.030
    Entpd5  Ectonucleoside triphosphate diphosphohydrolase 5  1  2.035  1.016  2.366  0.041
(b) Genes downregulated in cadmium-treated mice
    Sntg2  Syntrophin, gamma 2  1  0.400  1.000  0.400  0.027
    G6pc  Glucose-6-phosphatase, catalytic  1  0.152  0.547  0.169  0.041
The animals had obtained either mock s.c. injections (−Cd) or s.c. injections with 20 µmol/kg body weight CdSO4 (+Cd) 6 h before sacrificing them. The expression values for each gene are given as mean value of three animals per group, normalized to the mean value of group Mtf1loxP −Cd (relative activity).
aMean values of three independent Affymetrix probe sets.
bMean value of six independent Affymetrix probe sets.
cMean value of two independent Affymetrix probe sets.