An autonomous BMP2 regulatory element in mesenchymal cells

Abstract

BMP2 is a morphogen that controls mesenchymal cell differentiation and behavior. For example, BMP2 concentration controls the differentiation of mesenchymal precursors into myocytes, adipocytes, chondrocytes, and osteoblasts. Sequences within the 3'untranslated region (UTR) of the Bmp2 mRNA mediate a post-transcriptional block of protein synthesis. Interaction of cell and developmental stage-specific trans-regulatory factors with the 3'UTR is a nimble and versatile mechanism for modulating this potent morphogen in different cell types. We show here, that an ultra-conserved sequence in the 3'UTR functions independently of promoter, coding region, and 3'UTR context in primary and immortalized tissue culture cells and in transgenic mice. Our findings indicate that the ultra-conserved sequence is an autonomously functioning post-transcriptional element that may be used to modulate the level of BMP2 and other proteins while retaining tissue specific regulatory elements.

Fig. 1

Ultra-conserved sequence (UCS) mediated repression is context-independent in immortalized and primary mesenchymal cells. Luciferase reporter genes were driven by the ribosomal protein L10 (RPL10), Bmp2 (Bmp), or cytomegalovirus (CMV) promoter only upstream of the luciferase gene (RPL10-LUC or BmpLuc or CMVLUC). The effect of 1,254 nt of BMP2 sequence bearing the full-length 3′UTR with UCS inserted downstream of the luciferase open reading frame (RPL10-Luc-3′UTR) on luciferase activity was assessed in C3H10T½ pluripotent mesenchymal cells (A, n=6). The effect of the isolated UCS (BmpLucUCS, CMVLucUCS, +83 to 446 nt relative to the Bmp2 stop codon) was assessed in primary mouse calvarial cells (B, n=6). Relative reporter activity is shown±SEM. We previously demonstrated in C3H10T½ cells that BmpLucUCS and CMVLucUCS were 62% and 76% less active relative to the corresponding promoter-only plasmids [Devaney et al., 2009]. The Bmp2- and RPL10-driven plasmids bear the SV40 late poly(A) signal (SVpA). The RPL10-Luc-3′UTR plasmid also bears the full-length 3′UTR with the two natural Bmp2 poly(A) signals (B2pA). The CMV plasmids bear the bovine growth hormone (BGHpA) signal.

Fig. 2

Cre-recombinase mediated excision of the ultra-conserved sequence in mice. Diagrams of a reporter transgene with the murine Bmp2 promoter (nt −1,237 to 471) and 3′UTR regions (nt 9,392–11,604) flanking the lacZ gene. ▽ marks the loxP sites flanking the ultra-conserved sequence (UCS, nt 9,392–10,200). The approximate locations of the two natural Bmp2 poly(A) signals at 10,332 and 10,619 [Fritz et al., 2004; Liu et al., 2008] relative to the promoter (yielding 3′UTRs of ~870 and 1,175 nt, respectively) are indicated by arrows (). Primers that flanked the loxP sites were used for PCR to demonstrate complete excision of the ultra-conserved sequence. DNA from two Bmp2 transgenic littermates were amplified for 24 to 30 cycles as marked. The right series of reactions used DNA from a Cre-positive littermate. Direct sequencing and the size of the amplification products confirmed the predicted sequences of both the intact or recombined transgenes. PCR quantification and sequencing of the junctions between transgene copies indicate that three copies of the intact transgene inserted into the genome in a head to tail orientation. In +Cre mice, recombination yielded one copy of the recombined transgene at this site.

Fig. 3

The ultra-conserved sequence represses lacZ reporter transgene expression in mesenchymal cells of the aorta and coronary vasculature. A: The aorta and coronary vasculature are strongly βgal-stained in Cre-expressing (+Cre), but not non-expressing (no Cre), adult hearts (8 weeks). B,C: Sections of whole-mount βgal-stained Cre-expressing littermates (ages indicated in panels) were immuno-histochemically stained with an antibody against a smooth muscle actin (α-SMA, red). The merged images illustrate colocalization of βgal activity with α-SMA positive smooth muscle cells (SMC). The enlarged section from the no Cre aorta shown in (C) illustrates that a few blue nuclei (arrows) were present in no Cre mice bearing the intact transgene.

Fig. 4

A BAC reporter transgene whose βgeo transcript lacks the BMP2 3′UTR is ectopically expressed in the coronary vasculature. A: The diagram shows the BAC-based Bmp2 reporter gene generated previously [Chandler et al., 2007]. A strong SV40 poly(A) signal after the βgeo open reading frame prevents inclusion of the Bmp2 3′UTR with the ultra-conserved sequence in BAC-derived mRNAs. The approximate positions of the PCR primers used for the data shown in C are indicated. B: βgeo activity (blue) in micrographs of sections from BAC-bearing transgenic animals was observed in coronary artery cells. This pattern resembles the ectopic expression observed in +Cre mice (Fig. 3). C: RNA and genomic DNA was collected from mice harboring the BAC transgene and was subjected to reverse transcription-qPCR. The upstream primer hybridized to the βgeo coding sequence. The downstream primers hybridized to sequence just upstream (control SVpA) or downstream (past SVpA) of the SV40 poly(A) signal. The signals were normalized to that obtained from genomic DNA. The positive genomic control showed that each primer pair amplified efficiently. The negative control with reverse transcriptase omitted from the cDNA synthesis reaction (−RT) showed that the RNA was not contaminated with genomic DNA. Detection of PCR product only in the reaction using the control SVpA primer, but not the past SVpA primer, indicates that the βgeo transcript was efficiently truncated by the SV40 poly(A) signal.

Fig. 5

The ultra-conserved sequence (UCS) represses a vector encoding functional BMP2 protein. A: C3H10T½ cells were transfected with plasmids bearing the human BMP2 coding region with (CMV-BMP2-UCS) or without (CMV-BMP2) the UCS in the CMV-driven vector described in Figure 1. The amount of BMP2 protein in media conditioned for 2 or 5 days after transfection was measured by ELISA (R&D Systems). The average BMP2 concentration relative to the CMV-BMP2 vector is shown±SEM. Day 2, n=2; day 5, n=6. No BMP2 was detected in untreated or mock-transfected cells or in cells transfected with the empty vector. Cells were then grown for 14 days, stained for alkaline phosphatase positive nodules, and lightly stained with eosin. The graph shows the average number of nodules per square centimeter±SEM. No nodules were detected in cells transfected with the empty vector. B: Representative plates showing stained C3H10T½ cells transfected with the empty expression plasmid or plasmids bearing the human BMP2 coding region with or without the UCS as described in (A).