Allosteric Effects of Pit-1 DNA Sites on Long-Term Repression in Cell Type Specification Kathleen M. Scully, Eric M. Jacobson, Kristen Jepsen, Victoria Lunyak, Hector Viadiu, Catherine Carrière, David W. Rose, Farideh Hooshmand, Aneel K. Aggarwal, and Michael G. Rosenfeld

Transgenic mice were generated as previously described [E. B. Crenshaw III, K. Kalla, D. M. Simmons, L. W. Swanson, M. G. Rosenfeld, Genes Dev. 3, 959 (1989); S. A. Lira, E. B. Crenshaw III, C. K. Glass, L. W. Swanson, M. G. Rosenfeld, Proc. Natl. Acad. Sci. U.S.A. 85, 4755 (1988)] and identified both by Southern blot analysis and PCR. Immunohistochemical analysis was performed on pituitary cells dispersed and plated after collagenase treatment and fixation with formalin. Specific rabbit a-rPrl (Vector, 1:1000), monkey a-rGH (NIDDK National Hormone and Pituitary Program, 1:1000), mouse a-hGH antibody (Biomeda, 1:50); and fluorescein- and rhodamine-coupled secondary antibodies against rabbit, human, or mouse (1:100) were used. -161/-146 and Sp1 conserved regions of the rat GH promoter were replaced with the sequence 5´-TACGTCGACTGCCT-3´ and the T₃RE was replaced with the sequence 5´-TACGTCGACTGCCTGGCATCTAC-3´.

Supplemental Text Corresponding to (11)

A fragment of the Pit-1 gene encoding the POU domain (amino acids 128-273) was subcloned and expressed as we described previously [E. M. Jacobson, P. Li, A. Leon-del-Rio, M. G. Rosenfeld, A. K. Aggarwal, Genes Dev. 11, 198 (1997)]. The POU domain was cocrystallized with 28- and 30-bp DNA oligonucleotides containing the exact prolactin Prl-1P and growth hormone GH-1 sequences, respectively. Both cocrystals were obtained from solutions containing 200 to 275 mM triethyl ammonium phosphate (pH 4.0). The Prl-1P cocrystals belonged to space group P1 with unit cell dimensions of a = 43.2 Å, b = 55.1 Å, c = 57.0 Å, a = 89.5°, b = 71.6°, and g = 78.0°. The GH-1 cocrystals belonged to space group C2 with unit cell dimensions of a = 113.0 Å, b = 47.9 Å, c = 107.5 Å, a = 90°, b = 117.1°, and g = 90°. Iodinated and brominated derivatives were prepared for by substituting iodo- and bromouracils for thymine residues on the DNA. All x-ray data were measured from frozen cocrystals. The Prl-1P native data (3.05 Å, 83% complete) were measured at BNL (beamline X4A) on imaging plates (Rmerge of 0.052), and the GH-1 native (3 Å, 85% complete) and brominated (3 Å, 79% complete) data were measured at CHESS (beamline A1) on an ADSC CCD detector (Rmerge of 0.098 and 0.099, respectively). Data from the iodinated derivatives were measured at home on an Raxis IV imaging plate area detector mounted on a rotating anode x-ray generator. The HKL program package was used to integrate the reflections in all cases [Z. Otwinowski, W. Minor, Methods Enzymol. 276, 307 (1997)].

The structures of both complexes were solved by molecular replacement using the program AMoRe [J. Navaza, AMoRe: Molecular Replacement, E. J. Dodson, S. Gover, W. Wolf, Eds. (Science and Engineering Research Council, Daresbury Laboratory, Warrington, UK, 1992)]. The search models were constructed from the "prolactin-like" complex solved previously [E. M. Jacobson, P. Li, A. Leon-del-Rio, M. G. Rosenfeld, A. K. Aggarwal, Genes Dev. 11, 198 (1997)]. The Prl-1P search model, consisting of the DNA and the POUS domains gave a clear rotation-translation solution that was then fixed and used to find the translation solution for the homeodomains. Once all the subdomains were in place, the program X-PLOR [A. T. Brunger, X-PLOR Version 3.1, A System for X-ray Crystallography and NMR (Yale Univ. Press, New Haven, CT, 1993)] was used to initiate rigid body refinement (20 to 8 Å, followed by 8 to 3 Å), leaving aside 10% of the reflections for R_free calculations. We used the iodinated data to check the validity of the MR solution by calculating an anomalous difference Fourier map, which showed clear positions for the iodines using the MR phases. The refinement was continued to the highest resolution limit (3 Å) of the native data by iterative cycles of torsion angle and B _factor refinement followed by manual rebuilding of the protein and DNA using the program O [A. T. Jones, J. Y. Zou, S. W. Cowan, M. Kjeldgaard, Acta Crystallogr. A47, 110 (1991)]. The final R_factor is 0.239 (R_free of 0.312) for all data between 10 and 3.0 Å, using the program package CNS in later stages of refinement [A. T. Brunger et al., Acta Crystallogr. D54, 905 (1998)]. The root mean square deviations on bond lengths and bond angles are 0.008 Å and 1.7°, respectively. The structure includes Pit-1 residues 130-198, 216-273 for monomer 1; residues 130-197, 217-273 for monomer 2; 24 base pairs of the 28-bp duplex; and 15 water molecules. As in previous POU domain cocrystal structures [E. M. Jacobson, P. Li, A. Leon-del-Rio, M. G. Rosenfeld, A. K. Aggarwal, Genes Dev. 11, 198 (1997)], the linker segments connecting the POUS domain to the POUH domain are disordered. For the GH-1 complex, the DNA register was confirmed by anomalous difference Fourier maps calculated using brominated and iodinated data and the MR phases. Rigid body refinement (from 15 to 8 Å, followed by 8 to 4 Å) with the brominated data, leaving aside 10% of the reflections for R_free calculations, followed by iterative cycles of torsional and group B _factor refinement and manual rebuilding lowered the R_factor to 0.289 and the R_free to 0.36. The refinement was extended to the highest resolution limit (3 Å), first using the brominated data and then the native CCD data, to give a final R_factor of 0.258 (R_free of 0.334) for all data between 10 and 3 Å. The root mean square deviations on bond lengths and bond angles are 0.021 Å and 2.5°, respectively. The structure includes Pit-1 residues 130-198, 217-273 for monomer 1; residues 130-197, 217-273 for monomer 2; the entire 30-bp DNA duplex; and 12 water molecules.

Supplemental Text Corresponding to (15)

Single-cell nuclear microinjection assays and analysis were performed as previously described [K. Jepsen et al., Cell 102, 1 (2000)].

Supplemental Text Corresponding to (21)

For the chromatin immunoprecipitation assays, pituitary glands were isolated, fixed with 1% formaldehyde for 10 min at room temperature, and treated as previously described [A. Hecht, M. Grunstein, Methods Enzymol. 304, 389 (1999)] Cross-linked adducts were resuspended and sonicated, resulting in DNA fragments of ~500 bp selected to include the CAP site and GH regulating sequences. Immunoprecipitation was performed using antibody-coated tosylactivated Dynabeads M-280 (Dynal, Oslo). Protein-bound, immunoprecipitated DNA was dissolved in TE buffer and treated at 65°C overnight to reverse cross-links. Digestion buffer [100 mM NaCl, 10 mM tris-HCL (pH 8), 25 mM EDTA (pH 8), and 0.5% SDS] was added to the sample and incubated for 2 hours at 50°C with 0.1 mg/ml Proteinase K (Sigma). Following extraction and precipitation, 35 cycles of PCR were performed at an annealing temperature of 56°C. Primers used for the mouse GH promoter were Forward: 5´-GGGTGGTCTCTGTAGCTGAGATCTTGCG-3´; Reverse: 5´-CCCTACCTGTTT TCCTCGACCCCAAGGC-3´. a-Pit-1 and a-T₃Rb IgGs were used as previously described [K. Jepsen et al., Cell 102, 1 (2000)], and IgG against the largest subunit of Pol II was supplied by Santa Cruz Antibody. In the two-step ChIP assay, material from the first step was eluted by alteration of pH to allow a second round of ChIP with different IgGs.

Supplemental Text Corresponding to (22)

Sequences encoding amino acids 2053-2300 of N-CoR were placed in frame 3´ of an HA tag and SV40 nuclear localization signal with the rabbit 0.65 kb b-globin exon 1/exon 2 splice intron upstream and 0.63-kb hGH poly(A) sequence downstream. This transcript was put under control of 3 kb of rat prolactin 5´ flanking sequences to target expression selectively to lactotropes [E. B. Crenshaw III, K. Kalla, C. K. Glass, L. W. Swanson, M. G. Rosenfeld, Genes Dev. 3, 959 (1989)]. Double-label immunohistochemistry was performed on 9-mm sections of formalin and ethanol-fixed paraffin-embedded pituitaries using antisera from sheep a-rPrl (Biogenesis, 1:800), monkey a-rGH (NIDDK, 1:1000), rabbit a-rGH (Chemicon, 1:200), and mouse a-HA antibody (BabCo, 1:1000). Fluorescence-coupled secondary antibodies used were donkey a-sheep Alexa Fluor 488 (Molecular Probes, 1:400) and goat a-human Cy3 (Jackson ImmunoResearch, 1:500). Secondary antibodies for light field images were a-rabbit-coupled alkaline phosphatase (1:500) and a-mouse-coupled peroxidase (1:100). For fluorescent staining, sections were counterstained with bis-benzimide and imaged using deconvolution microscopy at ×40 magnification with a DeltaVision microscope (Applied Precision Instruments). Fifty 0.25-mm-spaced optical sections were imaged from each sample at multiple wavelengths, and individual optical sections were analyzed by SoftWorx (API).

To Advertise   |   Find Products