Identification of highly conserved nonexonic sequences in Agc1. (A) Diagram of the rat Agc1 gene showing the locations (in parentheses) and sizes of A1, A2, and A3. The horizontal line represents nonexonic sequences. Black boxes with numbers underneath represent exons. The arrow symbolizes the start of transcription, and AATAAA indicates the polyadenylation site. (B) Percentages of identity of A elements in mammalian genomes. ND, not determined. (C) ClustalW alignment of A1 sequences. The black histogram schematizes the degree of conservation of each nucleotide. The consensus sequence derived from the alignment is shown below the histogram. The L-Sox5/Sox6 and L-Sox5/Sox6/Sox9 binding sites identified in Fig. 7 are underlined with single-shaft and double-shaft arrows, respectively.

Agc1-lacZ expression in transgenic mice. (A) Schematic of Agc1-lacZ. Four tandem copies of A1 were inserted in an intron downstream of the −309 to +308 Col2a1 sequence and upstream of lacZ. The intron (bent line) is flanked with splice donor (SD) and acceptor (SA) sites. (B) Transient transfection of RCS cells to compare the activity of Agc1-lacZ to those of similar reporters featuring no enhancer (−) or the 4×48 Col2a1 enhancer. (C) X-Gal staining of wild-type (left) and Agc1-lacZ (right) littermates at embryonic day 13.5 (E13.5) and E15.5. (D) Frozen section through an E15.5 Agc1-lacZ embryo stained with X-Gal. CC, chondrocranium; N, nasal cartilage; M, Meckel's cartilage; S, sternal cartilage; T, tracheal cartilage; V, vertebral cartilage. (E) Paraffin sections of various skeletal elements in E14.5 Agc1-lacZ embryos stained with X-Gal. The digits, which are the least developed, show chondrocytes well differentiated and positive for X-Gal only in the cartilage core regions. The knee, which is more advanced, shows positive staining in most chondrocytes of the femur (F) and tibia (T) but not in the cruciate ligament (CL) and in other cells. The elbow, which is the most advanced, shows positive staining throughout humerus (H), ulna (U), and radius (R) cartilage. (F) Frozen sections of an E15.5 Agc1-lacZ embryo stained with X-Gal. (Left) Positive cells in the Meckel's cartilage but not in surrounding bone (B) and other tissues. (Middle) X-Gal-positive articular (AC), proliferating (PC), and hypertrophic (HC) chondrocytes in vertebral arches. (Right) X-Gal-positive cells in vertebral body (VB) cartilage and intervertebral disc nucleus pulposus (NP). (G) Frozen sections of the distal ilium growth plate of a newborn Agc1-lacZ mouse showing positive X-Gal staining in most chondrocytes. A few bone (B) and ligament (L) cells are X-Gal positive as a result of endogenous beta-galactosidase, as proven by the fact that the same result was obtained with nontransgenic littermates (data not shown). (H) Thoracic cage and knees of 2-week-old wild-type and Agc1-lacZ pups showing X-Gal staining derived from the lacZ transgene in the cartilaginous, ventral portion of the ribs (C) and endogenous beta-galactosidase activity in the ossified, dorsal portion of the ribs (B). F, femur; T, tibia. (I) Histology sections through a rib, vertebral joint, and knee of a 6-month-old Agc1-lacZ mouse, demonstrating positive X-Gal staining in articular (AC) and growth plate (GP) chondrocytes but not in bone marrow (BM).

A1 binding by the Sox trio in RCS cells. Chromatin immunoprecipitation in RCS and BALB/3T3 cells using Sox and control antibodies, as indicated. Shown are the products obtained upon PCR of immunoprecipitated chromatin with primers for the Agc1 and Col2a1 enhancers and a nonconserved control sequence located 2 kb downstream of A1. The sizes of the products are indicated. The first lane in each panel shows DNA markers, with their sizes indicated, and the second lane shows the PCR product obtained with chromatin input. n.i., nonimmune IgG. The data are representative of three experiments.

Delineation of Sox binding sites in A1. (A to F) EMSA of extracts from HEK-293 cells forced to express Sox6 or SOX9 incubated with probes carrying mutations, as indicated. Probes are named according to the positions of mutated nucleotides (A to D), number of nucleotides separating the Sox-like sites (E), and orientation of the sites (F). Shown are the Sox/DNA complexes. Arrows with single and double shafts designate SOX9 monomers and dimers, respectively (D and E). (G) Comparison of Sox9 binding sites in cartilage enhancers. The upper strand of the Sox consensus site is shown in reverse and forward orientations, as indicated with arrows and numbers. The Sox-like sites in enhancers are boxed, with boldface letters for consensus nucleotides and underlined letters for nucleotides present in most enhancers. (H) Comparison of L-Sox5/Sox6 binding sites in the Agc1 and Col2a1 enhancers. All 48 bp of the Col2a1 enhancer and flanking nucleotides are shown on two lanes.

Test of the importance of L-Sox5/Sox6 for Sox9 binding to cartilage enhancers. (A) ChIP in Sox5^fl/fl Sox6^fl/fl chondrocytes 48 h after infection with no adenovirus or AdeCre. Cartilage enhancers were detected by PCR of chromatin precipitated with antibodies as indicated. Binding of RNA polymerase II (Pol2) to the Gapdh promoter was used as positive control. (B) Quantitative ChIP assay of the same samples as in panel A, precipitated with IgG control (−) and Sox antibodies, as indicated. (C) Model for the mode of transactivation of Agc1 by the Sox trio. See text for details.

Enhancer activity of A1 in transiently transfected cells. (A) Schematics of reporter constructs. Each construct contains a −89 to +6 Col2a1 promoter (pCol2) or −596 to +90 Agc1 promoter (pAgc1) driving the firefly luciferase gene (Luc). Some constructs contain up to four tandem copies of the Col2a1 48-bp enhancer (4×48) or Agc1 A element (4×A). (B to F) Transient transfection of RCS, HEK-293, and BALB/3T3 cells with reporter constructs containing the indicated promoter and enhancer elements. Activation of the Col2a1 enhancer construct by SOX9 is shown as positive control in panels E and F.

A1 activation by the Sox trio. (A) Western blot of Sox5^fl/fl Sox6^fl/fl primary chondrocytes cultured for 48 h with no adenovirus (−), AdeGFP, or AdeCre. Primary antibodies and the M_rs of protein markers are indicated. (B) Beta-galactosidase activity in Sox5^fl/fl Sox6^fl/fl Agc1-lacZ primary chondrocytes cultured as described for panel A and harvested 24 h to 96 h after adenovirus infection. (C) Transfection of Cos-1 and HEK-293 cells with pCol2-Luc reporters harboring no enhancer (−), 4×48, or 4×A1 and Sox expression plasmids, as indicated. (D) Transfection of primary skin fibroblasts with pCol2-Luc reporters harboring no enhancer (−) or 4×A1 and Sox expression plasmids, as indicated. (E) Beta-galactosidase activity in Agc1-lacZ primary skin fibroblasts transfected for 36 h with Sox6 and/or SOX9 expression plasmids, as indicated.

Delineation of A1 regions bound by the Sox trio. (A) Schematic of EMSA probes. Probes A to H are about 60 bp and cover the entire A1 sequence, with 15-bp overlaps. Probes I to M are 23 to 40 bp. Stars represent the regions bound by L-Sox5/Sox6 and Sox9, as identified in panels B and C. (B) EMSA of RCS nuclear extracts with 2HMG and A to H probes and with antibodies, as indicated. The migration levels of free probes, antibody supershifts (Sox/AB), specific complexes (specific), and nonspecific complexes (non-sp) are indicated. (C) EMSA of HEK-293 and RCS extracts with 2HMG and I to M probes. HEK-293 cell extracts were prepared following transfection with Sox expression plasmids, as indicated.

Test of the functional importance of the Sox trio binding sites. (A) Transient transfection of RCS cells and HEK-293 cells with pCol2-Luc constructs containing two copies of A1 fragments. Fragments are schematized with numbers indicating nucleotide boundaries. The HEK-293 cells were cotransfected with Sox expression plasmids, as indicated. (B) Similar transfection as in panel A but with pCol2-Luc constructs harboring point mutations in the Sox sites of the 115 to 359 A1 fragment. The mutations corresponded to those described as m7-11-20-24 for segment I (Mut I), m5-9-12-16 for segment J (Mut J), m7-11-15-19-22-26 for segment K (Mut K), and m7-10-14-16 for segment L (Mut L) in Fig. 7.