HBP1
HBP1 is a single HMG-box transcriptional repressor and cell-cycle inhibitor, identified as a target of the retinoblastoma protein family. Its highest-affinity site, found in the human CD2 locus control region, is the twelve-base element TTCATTCATTCA, which is explicitly not the canonical HMG-box consensus used by SOX and TCF/LEF proteins; read on the other strand it is three tandem copies of TGAATGAA, the octamer that CIC recognizes. HBP1 represses the N-MYC and cyclin D1 promoters, opposes E2F at N-MYC, activates p16 and p21, and arrests the cell cycle even under strong growth signals. It is suppressed by AKT signaling and by several microRNAs, and behaves as a tumor suppressor without being a commonly mutated cancer gene.
The factors
One protein. HBP1 (HMG-box transcription factor 1) is a roughly 514-residue nuclear protein with a single HMG box. It was cloned twice independently: first by expression cloning in a potassium transport-defective yeast screen 7937077, and then as a repressor that binds the retinoblastoma family 9030690. It carries two RB-interaction motifs, LXCXE and IXCXE, and interacts selectively with RB and p130 but not p107; HBP1, RB and p130 all rise during differentiation while p107 falls 9030690. HBP1 is broadly expressed. Functionally it is a repressor and a brake on proliferation: it represses the N-MYC promoter and acts there as the opposite signal to E2F-1 9030690, represses CCND1 11500377, activates CDKN2A (p16) and CDKN1A (p21) 23249948, and its expression arrests the cell cycle even in the presence of optimal proliferation signals 9030690. Much of its effect is exerted through proteins rather than DNA - it neutralizes LEF1, TCF4 and MYC in the WNT/beta-catenin pathway 11500377,30683982. It sits downstream of growth factor signaling in two ways: AKT phosphorylates HBP1 and inhibits its transcriptional activity 30683982, and the PI3K/PKB/FOXO pathway controls HBP1 transcription by regulating FOXO binding to the HBP1 promoter 24762137.
The motif
The consensus given here, TTCATTCATTCA, is the highest-affinity HBP1 binding sequence described. It was found in HSS3 of the human CD2 locus control region and binds more tightly than the HBP1 sites reported earlier in cell-cycle promoters 10562551. The authors noted explicitly that it differs from the canonical (T/A)(T/A)CAA(A/T)G consensus of other HMG-box proteins 10562551,8190643, so HBP1 should not be scanned with a SOX or TCF/LEF model. The element is a tandem trimer: read on the opposite strand it is TGAATGAATGAA, three copies of the octamer TGAATGAA that CIC binds, and both proteins are HMG-box repressors, which makes sequence-only discrimination between them impossible at the level of a single octamer. Length matters a great deal here. An exact twelve- base match is expected only a few hundred times in a 3.1 Gb double-stranded genome, which makes a full match informative; a single TTCATTCA octamer is expected of order a hundred thousand times and is not. The motif model carried in these databases is of no use for scanning: the sole entry, CIS-BP M00194_2.00, derived from mouse Hbp1 at full DNA-binding-domain identity, contains 1.1 bits of information in total 25215497, and its consensus string collapses to a single degenerate base. Species is not the problem - the DNA-binding domains are identical - the problem is that the model records almost no preference. Its best similarity to any motif outside the family is 0.797, and being a one-member family it has no within-family comparison.
Where it acts
The place HBP1 binding was best characterized is unusual. HSS3 of the hCD2 locus control region has no enhancer activity of its own but is required to prevent position-effect variegation: transgenes carrying an HSS3 deletion of the HBP1 sites variegate when they integrate into centromeric or telomeric heterochromatin 10562551. The interpretation offered was architectural - HBP1 binds and bends the DNA, making protein-DNA and protein-protein contacts more likely and raising the probability that an active locus is established 10562551. Elsewhere HBP1 is documented at cell-cycle gene promoters, including N-MYC, CCND1, CDKN2A, CDKN1A and DNMT1 9030690,11500377,23249948. There is no genome-scale map of HBP1 occupancy in these collections, and no HBP1 model from ChIP-seq in any of the four databases.
Biology and disease
HBP1 behaves as a cell-cycle inhibitor and candidate tumor suppressor. It contributes to oncogene-induced senescence downstream of RAS and to terminal differentiation, and its repression of DNMT1 links it to the senescence program 23249948. It is not among the frequently mutated cancer driver genes, but its expression is lower in several tumor types than in matched normal tissue, and alterations of HBP1 have been reported in invasive breast cancer 17616670. Loss of HBP1 expression is more often post-transcriptional or signaling-driven than mutational: miR-155, miR-17-92 and miR-29a each reduce HBP1 levels in different tumor contexts 30683982, and growth factor signaling through PI3K/PKB/FOXO downregulates HBP1 transcription while AKT inactivates the protein directly 24762137,30683982.
Reading a hit
Do not use the database model for this family; at 1.1 bits it will match essentially anywhere. If you want HBP1 candidate sites, scan the published element TTCATTCATTCA, and treat a full twelve-base match as meaningful and a single octamer as not. The main interpretive hazard is CIC: the octamer unit is identical, so a TTCATTCA or TGAATGAA hit is compatible with either protein, and expression data or the presence of a tandem trimer versus a lone octamer is what separates them. Do not read a SOX- or TCF/LEF-type WWCAAWG match as an HBP1 site; HBP1 was explicitly shown not to use that consensus 10562551. Because there is no ChIP-derived HBP1 model here, nothing in this entry is contaminated by indirect binding - but equally, nothing has been tested genome-wide. Finally, remember that a large part of HBP1's activity is protein-protein: absence of a site near a gene HBP1 regulates does not mean HBP1 is not involved.
How this was worked out
HBP1 is cloned from rat by its ability to suppress a potassium transport defect in yeast, turns out to encode a 513-residue HMG-box protein unrelated to any transporter, and is found to rise with differentiation in adipocyte and myogenic cell lines.
Still open then: Neither a DNA-binding site nor a transcriptional function had been demonstrated.
HBP1 is isolated as a target of the retinoblastoma family: it carries LXCXE and IXCXE motifs, binds RB and p130 but not p107, represses the N-MYC promoter in opposition to E2F-1, and its expression arrests the cell cycle despite optimal growth signals.
HBP1 is shown to separate cell-cycle exit from differentiation: in C2C12 myoblasts it blocks MyoD-family-driven muscle differentiation while leaving cell-cycle exit intact, with the outcome set by the RB/HBP1 ratio.
The highest-affinity HBP1 site is identified as TTCATTCATTCA in HSS3 of the human CD2 locus control region, explicitly different from the canonical HMG-box consensus; deleting it makes transgenes variegate when they land in heterochromatin.
Still open then: Whether this element is used at HBP1's cell-cycle target promoters, or anywhere else in the genome, was not tested.
HBP1 is shown to inhibit Wnt signaling by binding TCF/LEF proteins and blocking their DNA binding, and to repress cyclin D1: a large part of its activity does not require an HBP1 site at all.
During senescence HBP1 represses DNMT1 through a site at -115 to -134 of its promoter while activating p16, and the two opposite activities are shown to proceed through different DNA sequences and different protein acetylation states.
Still open then: Leaves unclear what single sequence, if any, defines an HBP1 site.
- 2014mediumMicroarray meta-analysis, reporter assays, chromatin immunoprecipitation, knockdown24762137
HBP1 is placed downstream of growth factor signaling as a FOXO target: FOXO binds a conserved site in the HBP1 promoter, PDGF and FGF lower HBP1 mRNA through PI3K and PKB, and HBP1 knockdown increases fibroblast proliferation.
A synthesis of the field argues that HBP1 behaves as a tumor suppressor without being a recurrently mutated driver, lost instead through microRNAs such as miR-155, miR-17-92 and miR-29a and through direct AKT phosphorylation.
Still unresolved
- Whether TTCATTCATTCA is the sequence HBP1 uses generally is unresolved: it was defined at one locus, the cell-cycle promoters HBP1 regulates were mapped to different and shorter sequences, and the repression and activation functions were shown to work through different DNA sequences, so no single element has been validated across targets.
- Whether HBP1 has any strong intrinsic sequence preference is unclear: the published CD2 element is a high-affinity site, yet the only in vitro model in these databases carries 1.1 bits, and no modern high-throughput measurement of HBP1 specificity has been published to settle whether the weak model reflects the protein or a failed assay.
- Nothing distinguishes an HBP1 site from a CIC site by sequence: the HBP1 element read on the other strand is three copies of the CIC octamer TGAATGAA, no genome-wide occupancy map exists for either protein in these collections, and whether the two repressors compete at shared sequences has not been examined.
- How much of HBP1's regulatory output requires DNA binding is open: it represses cyclin D1 and Wnt targets by sequestering LEF1, TCF4 and MYC, so the relative contributions of site occupancy and protein-protein antagonism at any given target gene are unmeasured.
Low confidence in any motif-based call. The single database model for HBP1 carries 1.1 bits and is unusable for scanning; the entry here gives the published high-affinity site instead, which rests mainly on one study of the human CD2 locus control region. The octamer unit of that site is identical to the CIC recognition octamer read on the other strand, so single-octamer matches cannot be assigned to HBP1 rather than CIC. A full twelve-base match is rare enough to be informative; a partial one is not. HBP1 also works substantially through protein-protein antagonism of LEF1, TCF4 and MYC, so its regulatory targets need not contain an HBP1 site at all.
References
- Lesage F, Hugnot JP, Amri EZ, et al. Expression cloning in K+ transport defective yeast and distribution of HBP1, a new putative HMG transcriptional regulator. Nucleic Acids Res 1994;22:3685-8. PMID 7937077
- Tevosian SG, Shih HH, Mendelson KG, et al. HBP1: a HMG box transcriptional repressor that is targeted by the retinoblastoma family. Genes Dev 1997;11:383-96. PMID 9030690
- Shih HH, Tevosian SG, Yee AS. Regulation of differentiation by HBP1, a target of the retinoblastoma protein. Mol Cell Biol 1998;18:4732-43. PMID 9671483
- Zhuma T, Tyrrell R, Sekkali B, et al. Human HMG box transcription factor HBP1: a role in hCD2 LCR function. EMBO J 1999;18:6396-406. PMID 10562551
- Sampson EM, Haque ZK, Ku MC, et al. Negative regulation of the Wnt-beta-catenin pathway by the transcriptional repressor HBP1. EMBO J 2001;20:4500-11. PMID 11500377
- Paulson KE, Rieger-Christ K, McDevitt MA, et al. Alterations of the HBP1 transcriptional repressor are associated with invasive breast cancer. Cancer Res 2007;67:6136-45. PMID 17616670
- Pan K, Chen Y, Roth M, et al. HBP1-mediated transcriptional regulation of DNA methyltransferase 1 and its impact on cell senescence. Mol Cell Biol 2013;33:887-903. PMID 23249948
- Coomans de Brachène A, Bollaert E, Eijkelenboom A, et al. The expression of the tumour suppressor HBP1 is down-regulated by growth factors via the PI3K/PKB/FOXO pathway. Biochem J 2014;460:25-34. PMID 24762137
- Bollaert E, de Rocca Serra A, Demoulin JB. The HMG box transcription factor HBP1: a cell cycle inhibitor at the crossroads of cancer signaling pathways. Cell Mol Life Sci 2019;76:1529-1539. PMID 30683982
- Harley VR, Lovell-Badge R, Goodfellow PN. Definition of a consensus DNA binding site for SRY. Nucleic Acids Res 1994;22:1500-1. PMID 8190643
- Weirauch MT, Yang A, Albu M, et al. Determination and inference of eukaryotic transcription factor sequence specificity. Cell 2014;158:1431-1443. PMID 25215497
1 motifs in this family — CIS-BP 1
| motif | database | consensus | bits |
|---|---|---|---|
| M00194_2.00Mus musculus 1 | CIS-BP | d | 1.1 |
HOCOMOCO tags read as assay + quality grade: P ChIP-seq, S HT-SELEX, M methyl-HT-SELEX, G GHT-SELEX, I SMiLE-seq, B PBM; grade A means two or more assay types agreed. A P-only matrix is a record of where the protein was crosslinked, which need not be a sequence it binds itself.
The factors, elsewhere
These entries are written from the primary literature and every citation is checked against PubMed, but this one can still be out of date, too confident, or simply wrong — especially where a motif’s published site and its database matrices disagree. Corrections and additions are welcome, and they are the fastest way to make this page better for everyone.