CBFβ is a facultative Runx partner in the sea urchin embryo
© Robertson et al. 2006
Received: 16 November 2005
Accepted: 09 February 2006
Published: 09 February 2006
Skip to main content
© Robertson et al. 2006
Received: 16 November 2005
Accepted: 09 February 2006
Published: 09 February 2006
Runx proteins are developmentally important metazoan transcription factors that form a heterodimeric complex with the non-homologous protein Core Binding Factor β (CBFβ). CBFβ allosterically enhances Runx DNA binding but does not bind DNA itself. We report the initial characterization of SpCBFβ, the heterodimeric partner of SpRunt-1 from the sea urchin Stronylocentrotus purpuratus.
SpCBFβ is remarkably similar to its mammalian homologues, and like them it enhances the DNA binding of the Runt domain. SpCBFβ is entirely of zygotic provenance and its expression is similar that of SpRunt-1, accumulating globally at late blastula stage then later localizing to endoderm and oral ectoderm. Unlike SpRunt-1, however, SpCBFβ is enriched in the endodermal mid- and hindgut of the pluteus larva, and is not highly expressed in the foregut and ciliated band. We showed previously that morpholino antisense-mediated knockdown of SpRunt-1 leads to differentiation defects, as well as to extensive post-blastula stage apoptosis caused by under-expression of the Runx target gene SpPKC1. In contrast, we show here that knockdown of SpCBFβ does not negatively impact cell survival or SpPKC1 expression, although it does lead to differentiation defects similar to those associated with SpRunt-1 deficiency. Moreover, SpRunt-1 containing a single amino acid substitution that abolishes its ability to interact with SpCBFβ retains the ability to rescue cell survival in SpRunt-1 morphant embryos. Chromatin immunoprecipitation shows that while the CyIIIa promoter engages both proteins, the SpPKC1 promoter only engages SpRunt-1.
SpCBFβ is a facultative Runx partner that appears to be required specifically for cell differentiation.
Runx proteins are transcription factors that function critically in gene regulatory networks that control cell proliferation and differentiation during animal development . The Runx family is defined by a highly conserved 128 amino acid sequence known as the Runt domain, which binds specifically to the DNA sequence TGT/CGGT [2, 3]. The sea urchin Runx transcription factor SpRunt-1 was discovered biochemically as a protein that binds this sequence in the cis-regulatory domain of the aboral ectoderm differentiation gene CyIIIa, an interaction that contributes to transcriptional activation of CyIIIa [4, 5]. SpRunt-1 transcripts are globally expressed in the early embryo, and then become localized predominantly to regions of continued growth and cell proliferation within larval oral ectoderm and endomesoderm . Morpholino antisense-mediated knockdown of SpRunt-1 has shown that it is required throughout the embryo for cell differentiation  and survival . The latter function is mediated by a positively-acting interaction between SpRunt-1 and its target gene SpPKC1, which encodes a conventional protein kinase C .
Runx proteins heterodimerize with another protein commonly referred to as core binding factor beta (CBFβ), a non-DNA binding protein that allosterically enhances the DNA binding of the Runt domain . Mammalian CBFβ also increases the half-life of Runx1 by protecting it from ubiquitin-mediated proteolysis . Since Runx1 contributes to cell cycle control and its protein levels are cell cycle regulated , this suggests that CBFβ may be a regulatory subunit that modulates Runx activity during the developmental transition from proliferation to terminal differentiation.
Knockout of CBFb in mouse leads to embryonic lethality caused by a failure of hematopoiesis [12, 13], a phenotype very similar to that caused by knockout of Runx1. As with RUNX1, chromosomal translocations involving CBFb are commonly associated with human leukemia . Like Runx2, CBFb has also been shown to be important for osteogenesis, although its loss-of-function phenotype is not identical to that caused by Runx2 knockout since some osteoblast differentiation does occur [14, 15]. Drosophila has two CBFβ genes, brother and big brother, the products of which contribute redundantly to the functionality of the Runx proteins Runt and Lozenge . These studies all indicate that CBFβ is important for Runx function; however, to date, no studies have definitively addressed the question of whether the Runx-CBFβ interaction is constitutive (obligate or context-independent) or facultative (regulatory or context dependent). The latter possibility is suggested by the facts that (a) the interaction is inhibited by sequences within mammalian Runx2  and the C-terminal domains of some Runx1 isoforms , and (b) Runx proteins can bind DNA as a monomer, albeit less effectively than as a heterodimer . Mouse CBFβ, which is ubiquitously expressed, has been shown to interact specifically with another protein termed Crl-1, which is expressed in subsets of neuronal cells . Finally, avian CBFb displays a complex pattern of expression in early development that is not identical to the pattern of Runx1 expression, with some regions of the embryo expressing only one or the other gene . Taken together, these observations invite speculation that Runx proteins might interact with only a subset of target genes (or with some target genes only part of the time) as a heterodimer with CBFβ, and that the choice between these alternative states may in part be determined by the specific Runx isoform as well as protein-protein interactions specified by the context of each particular tissue, cell type, and cis-regulatory system.
Here we describe the cloning and initial characterization of SpCBFβ, the heterodimeric Runx partner from the sea urchin Strongylocentrotus purpuratus. We show that SpCBFβ is expressed in a pattern that is similar but not identical to that of SpRunt-1. Moreover, while SpCBFβ participates in the Runx-dependent activation of several genes including CyIIIa, it is not required for or involved in Runx-dependent activation of SpPKC1, which we showed previously to be critical for cell survival in the embryo . Therefore, heterodimerization of SpRunt-1 with SpCBFβ occurs facultatively in the sea urchin, and is a context-dependent aspect of Runx-mediated transcriptional control.
Electrophoretic mobility shift analysis (EMSA) was used to test whether SpCBFβ enhances the DNA binding activity of SpRunt-1. Recombinant SpRunt-1 protein was reacted with a fluorescently labeled oligonucleotide probe containing the SpRunt-1 binding site from CyIIIa , either with or without pre-incubation with recombinant SpCBFβ. DNA binding of full-length SpRunt-1, or a fragment containing only the Runt domain thereof, is substantially enhanced in the presence of SpCBFβ, which forms a complex with the SpRunt-1 protein (as indicated by a "supershift"; Fig. 1B, lanes 2 and 4), whereas SpCBFβ alone does not bind DNA (Fig. 1B, lane 5). It is likely that native SpRunt-1 forms a heterodimeric complex with SpCBFβ in the embryo, and consistent with this, affinity purified SpRunt-1 protein from blastula stage nuclear extracts was initially identified as a heterodimer containing a 21 kDa subunit . To confirm this, an antibody generated against recombinant SpCBFβ was used to probe an immunoblot of whole nuclear extract, nuclear extract immunodepleted with an anti-SpRunt-1 antibody, and the resulting immunoprecipitate. A 21 kDa SpCBFβ-immunoreactive band present in blastula stage nuclear extract (Fig. 1C, lane 1) is specifically immunodepleted from the extract by the SpRunt-1 antibody (Fig. 1C, lane 2), and pulled down in the SpRunt-1 immuoprecipitate (Fig. 1C, lane 3), whereas the band is not similarly precipitated by non-specific IgG (Fig. 1C, lanes 4 and 5). These data suggest that SpCBFβ exists in a complex with SpRunt-1 in vivo.
Quantitative RT-PCR was also used to measure the abundance of SpCBFβ transcripts. Toward this end a standard curve was constructed from in vitro synthesized SpCBFβ mRNA (see Methods). We used the same method to measure the abundance of SpRunt-1 transcripts, which were previously measured by an RNAase protection method . The measurements of SpRunt-1 were in agreement with our previous measurements, indicating 7,000–8,000 transcripts per blastula or gastrula stage embryo. In the mid-gastrula stage embryo, when SpCBFβ is maximally expressed, it is present at slightly higher levels than is SpRunt-1, determined by qRT-PCR to be 12,000 transcripts per embryo. Thereafter, SpCBFβ transcripts decline in abundance to levels approximating those of SpRunt-1.
SpRunt-1 derived from exogenous mRNA rescues cell survival in SpRunt-1 morphants . If SpCBFβ is dispensable for the survival function of SpRunt-1, then a SpRunt-1 protein that has been mutated so that it cannot interact with SpCBFβ should retain the ability to rescue cell survival in SpRunt-1 morphants. To test this, we constructed a SpRunt-1 mutant protein wherein an arginine is substituted for a glycine residue that is essential for the interaction between the Runt domain and CBFβ (SpRunt-1-G115R). Electrophoretic mobility shift analysis verifies that the Runt domain from SpRunt-1-G115R retains the ability to bind DNA as a monomer, but does not heterodimerize with SpCBFβ (Fig. 5C). Importantly, SpRunt-1-G115R rescues cell survival in SpRunt-1 morphants (Fig. 5D), indicating that heterodimerization with SpCBFβ is not required for the anti-apoptotic function of SpRunt-1.
While these results indicate that SpCBFβ is not required for the transcriptional activation of SpPKC1, they do not rule out the possibility that SpCBFβ nonetheless heterodimerizes with SpRunt-1 on its target sequences in the SpPKC1 promoter. To address this issue, we performed chromatin immunoprecipitation (ChIP) of chromatin prepared from late gastrula stage embryos using antibodies to either SpRunt-1 or SpCBFβ. Consistent with the expression data indicating that CyIIIa activation requires the heterodimer, sequences from the CyIIIa promoter are recovered by ChIP using both antibodies, as are sequences from SpCyclinD (Fig. 6B). In contrast, whereas sequences from the SpPKC1 promoter are recovered by ChIP using the SpRunt-1 antibody, they are not recovered by the SpCBFβ antibody (Fig. 6B). We conclude that in the gastrula stage embryo, SpCBFβ does not exist in a heterodimeric complex with SpRunt-1 within the context of the SpPKC1 promoter.
Our data are consistent with the proposition that CBFβ is a facultative Runx partner that participates in the regulation of a subset of Runx target genes. It will be important to analyze the SpPKC1 cis-regulatory system to learn the contextual rules that facilitate CBFβ-independent Runx function, and to determine the biological rationale for Runx-mediated transcriptional regulation that does not involve CBFβ. One intriguing possibility is that genes that need to be rapidly responsive to physiological signals require Runx-DNA binding complexes that are less stable and/or have shorter half lives. Addressing this issue will require identification and comparison of the cis-regulatory systems from additional CFBβ-dependent and -independent Runx target genes.
Adult Strongylocentrotus purpuratus were obtained from Charles Hollahan (Santa Barbara Marine Biologicals, Santa Barbara, CA) or Pat Leahy (Corona del Mar, CA). Gamete collection and embryo culture were carried out as described previously .
Embryo microinjections and imaging were carried out as previously described .
An alignment of all the known CBFβ protein sequences was used to design a degenerate PCR primer set. The following oligonucleotide primers were chosen to target conserved regions of the sequence alignment having amino acids with low codon degeneracy: N-terminal region MPRVVPDQ giving forward primer ATG CCI MGI GTI GTI CCI GAY CA, and central region NGVCV(LIR)W(RK)GW giving reverse primer CCA GGI ICK CCA IMD IAC RCA IAC ICC RTT (where M = A+C/Y = C+T/K = T+G/D = A+T+G/R = A+G). The SpCBFβ message was amplified from total RNA isolated from a 12 hour S. purpuratus embryo culture using 2-step PCR. Random primer was used in the cDNA synthesis step followed by the use of degenerate primers to direct PCR amplification of SpCBFβ from the random primed template. The single 350 bp PCR product, which was the size expected for a conserved sequence relative to the consensus, was desalted and cloned into plasmid vector pGEM T easy® (Promega) by TA cloning. A set of 6 separate clones were fully sequenced and theoretical translation showed clear homology to consensus CBFβ. The nucleotide sequence of the partial SpCBFβ clone obtained by PCR with degenerate primers was in turn used to design specific primers for RACE. 5' RACE was performed starting with total RNA isolated from a 24 hour S. purpuratus embryo culture. Using GeneRacer™ (Invitrogen) following the manufacturer's protocol with the random primer option, 5' RACE cDNA was synthesized and used as a template for PCR with the included 5' RACE primer as a forward primer and a SpCBFβ-specific reverse primer. For 3' RACE, the phosphatase and ligation steps described in the GeneRacer™ protocol, designed to enrich for the 5' end, were eliminated. 3' RACE cDNA was synthesized with the oligo dT primer option and used as a template for PCR with an SpCBFβ specific forward primer, the included 3' RACE primer serving as the reverse primer. The final SpCBFβ sequence (Genbank accession number DQ205186) could not be extended further by RACE or database searches and is considered to be full-length.
Full length SpRunt-1 and the Runt domain thereof were each subcloned from the original pSPORT1 plasmid into the bacterial expression vector pRSET A (Invitrogen). For subcloning, PCR primers were designed to insert the full length or partial coding sequence in frame relative to the his-tag encoded in the vector sequence. Likewise, the full coding region of SpCBFβ was subcloned into pRSET A downstream of the his-tag. For the SpRunt-1-G115R, base substitutions were made in the SpRunt-1 Runt domain/pRSET clone by the QuickChange® II Site-Directed Mutagenesis method (Stratagene), using the following primers: Forward T ACG ATG GTG ACG ATC GCA GCA GGG AAC GAT GAA AAC T and Reverse T ACG ATG GTG ACG ATC GCA GCG CGC AAC GAT GAA AAC T. The based substitutions introduced a unique BssHll site for diagnostic purposes.
Each of the expression constructs was induced in a BL21-pLysS culture, and protein was isolated from the IPTG-induced culture using a Ni-NTA Spin Kit (Qiagen). EMSA was performed with fluorescently labeled oligonucleotides as previously described , using either the CyIIIa Runx site described previously  or the following Runx-site oligonucleotides as probe (from the SpCyclinD promoter region): 5'-/5Cy5/ATT ATT CTC TGA CCA CAA TTT TTG TTA GA-3' and 5'-TCT AAC AAA AAT TGT GGT CAG AGA ATA AT-3'.
RNA blot analysis was carried out on total RNA from staged embryos as described previously .
For quantitation of relative transcript levels in control versus knockdown embryos, RT-PCR was performed as described previously , using RNA prepared from either 48 h (for SpPKC1, CyIIIa, and SpDri) or 36 h (for SpCyclinD) embryos. For quantitation of absolute transcript numbers, a SpCBFβ standard was prepared using the T7 mMessage Machine (Ambion) with the linearized clone as an in-vitro transcription template. After quantitation the SpCBFβ transcript was diluted in a 10 fold series ranging from 2.5 × 103 to 2.5 × 108 transcript copies and each dilution was adjusted to a final 5 μg amount by the addition of total RNA extracted from eggs (there is no endogenous SpCBFβ transcript detectable in eggs). Each of the diluted standards as well as 5 μg 24 and 40 hour total RNA samples were amplified by PCR using a 2-step procedure. First strand synthesis was performed using Superscript™ lll (Invitrogen) with the random primer option. PCR was then performed in a SmartCycler II (Cepheid) with SpCBFβ-specific primers and SYBR green detection. The final standard curve was used along with known values for total RNA yield per embryo and known number of embryos used for the sample RNA extraction to determine the transcript copy number per embryo.
The SpRunt-1 polyclonal antibody was described previously . The SpCBFβ polyclonal antibody was similarly produced in rabbits by Cocalico Biologicals (Reamstown, PA), using purified recombinant SpCBFβ as antigen. The SLBP antibody was a gift of Dr. William Marzluff (University of North Carolina). Immunoblots were performed as described previously .
Morpholino antisense oligonucleotides (MASOs) were designed by and purchased from GeneTools, LLC (Corvallis, OR). The sequence of the SpCBFβ MASO is: CTACTCTGGGCATAGTTGACATCGG. The SpRunt-1 MASO m5 was described previously . The standard non-specific control MASO from GeneTools was used as a negative control.
TUNEL assays on fixed embryos to image apoptosis were performed as previously described .
For rescue experiments, a mutation that changed Gly 115 to Arg was made in the SpRunt-1 full length rescue clone  by the QuickChange® II Site-Directed Mutagenesis method (Stratagene), using the primers described above for the SpRunt-1 expression plasmid. The T7 mMessage Machine (Ambion) was used to synthesize full-length capped mRNA for microinjection.
ChIP was carried out as previously described , using chromatin prepared from 48 h embryos and polyclonal IgG specific for SpRunt-1 and SpCBFβ. Non-immune or pre-immune IgG was used as a specificity control for anti-SpRunt-1 and anti-SpCBFβ antibodies, respectively. The following primers were used for PCR amplification ChIP samples: SpPKC1: GACCCCTGGCTTAATATGTTGATGTGTT (forward) and CCTTCATCTCAAACGAAGAATCCGACAT (reverse); CyIIIa: GTAGCACACGGAGAGATTGTGGGACAT (forward) and GGATCGGGGTTAGAGTTACATTTGGCTT (reverse); SpCyclinD: AGAAACGAATGTATCCGTGTGTTGTGAA (forward) and GCGAGACATAACTTCCTTGATCGTGCTA (reverse); and SpCdk4: CAGGAGCGTAGTCAATCCGCATCAA (forward) and CAGCCTGCAACTTCTGAGATGCTTTGT (reverse).
We thank Dr. Eric Davidson for reviewing this manuscript prior to submission, the four anonymous reviewers for helpful comments that further improved the manuscript, and the Molecular Biology Core of the Stowers Institute for DNA sequencing. This work was funded by the Stowers Institute for Medical Research and by a grant from the NIH (GM070840).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.