NR4A (NUR77/NURR1)
NR4A1 (NUR77), NR4A2 (NURR1) and NR4A3 (NOR-1) are orphan nuclear receptors and immediate-early genes: they are transcribed within minutes of serum stimulation, depolarization or antigen-receptor signaling and their proteins turn over quickly. Unlike most nuclear receptors they bind DNA as monomers, to a single AGGTCA half-site carried on an A/T-rich 5' extension, giving the eight-base NGFI-B response element AAAGGTCA. A separate protein surface, the A box, reads those A-T base pairs in the minor groove. The same receptors can also act as homodimers on the everted NurRE and as heterodimers with RXR. Because a monomeric half-site constrains only eight or nine positions, matches are common by chance, and the AGGTCA core is shared with most of the nuclear receptor superfamily.
The factors
The NR4A subfamily has three paralogues: NR4A1 (NUR77, NGFI-B, TR3), NR4A2 (NURR1) and NR4A3 (NOR-1, MINOR). Each has the standard nuclear receptor layout: an AF-1 region, a DNA-binding domain of two C4 zinc modules, a carboxy-terminal extension containing the A box, and a ligand-binding domain. All three are immediate-early genes induced by serum, growth factors, depolarization, antigen-receptor signaling and stress, with short protein half-lives, so levels are low in resting tissue; NR4A1 was cloned as a nerve growth factor-induced transcript. They are orphans in a strong sense: the NURR1 ligand-binding domain has no accessible ligand pocket, its interior filled with bulky side chains, defining a class of receptors that activate without a ligand (Wang et al. 2003) 12774125. The three paralogues are largely redundant: NR4A1-null mice have normal thymocyte apoptosis (Lee et al. 1995) 7624775. Ten motifs are collected here across all four sources.
The motif
The core element is the NGFI-B response element, NBRE, AAAGGTCA, identified by genetic selection in yeast (Wilson et al. 1991) 1925541. Its structure explains its length. The two zinc modules read the AGGTCA half-site through the major groove as in other nuclear receptors, while a substructure carboxy-terminal to them, the A box, reaches into the minor groove and reads the three A-T base pairs extending 5' of the half-site; this was the first structure of a nuclear receptor bound to DNA as a monomer (Meinke and Sigler 1999) 10331876. Domain swaps had already shown that NR4A1 and SF-1 (NR5A1) bind by the same monomeric mechanism with distinct 5' requirements: an amino group in the minor groove of the SF-1 sequence and an asparagine in the SF-1 A box (Wilson et al. 1993) 8395013. The same receptors use two other arrangements: homodimers on the NurRE, an everted pair of NBRE-related octamers with an unusual spacer, which mediates signaling at the POMC promoter (Philips et al. 1997) 9315652, and heterodimers with RXR on a direct repeat (Perlmann and Jansson 1995) 7705655. The collected data agree: the pooled consensus is AAAGGTCA, and five models from three databases give exactly that string. The highest-information model, the Vierstra archetype AC0119:NR4A at 25 bits, is GACCTTTAAAGGTCA: two NBRE octamers head to head with no spacer, not the classic NurRE spacing. Two entries need care. JASPAR NR4A2::RXRA (MA1147.2) is an inverted-repeat dimer model whose nearest neighbors sit outside this family, so it describes a heterodimer, not monomer binding. The archetype AC0114:NR4A/NR2C is flagged in the data as probably belonging with the RXR heterodimers: it is a bare aAGGTCA half-site, exactly what one expects to be ambiguous. Nothing distinguishes the three paralogues by sequence.
Where it acts
NR4A sites behave like those of other stimulus-induced factors: promoters and signal-responsive enhancers, with occupancy tracking induction of the receptor rather than presence of the motif. They act in T cells, neurons, pituitary corticotropes, macrophages, vascular smooth muscle, liver and adipose tissue, so target programs are cell-type specific. A specific eight-mer is expected about 47,000 times per strand in a 3-Gb genome, so tens of thousands of NBRE matches exist and only a small fraction is ever bound. Composite arrangement matters: NurRE-type pairs and NBRE-RXR direct repeats bind more tightly and are the configurations shown to respond to physiological signals, while a lone NBRE is often weak or non-functional.
Biology and disease
NR4A2 loss abolishes midbrain dopaminergic neurons in mice and the gene has been examined in Parkinson disease (Zetterstrom et al. 1997) 9092472. Although single knockouts are near-normal, the subfamily as a group limits T cell effector function: deleting all three NR4A genes relieves an exhaustion-like program and improves tumor control by CAR T cells (Chen et al. 2019) 30814732. In the pituitary, NR4A1 mediates CRH-induced POMC transcription through the NurRE (Philips et al. 1997) 9315652. The clearest cancer link is structural: extraskeletal myxoid chondrosarcoma is defined by a translocation fusing EWSR1 to NR4A3, converting the receptor into a constitutive activator (Labelle et al. 1995) 8634690.
Reading a hit
A monomeric half-site is much harder to call than a spaced dimer; this is the central practical point. A dimeric element such as a DR3, DR4 or DR5 constrains two 6-bp half-sites plus a fixed spacer, 13-15 positions in all, so a scanner can demand a rare pattern and the arrangement itself is evidence. The NBRE constrains only eight or nine positions, of which six are the AGGTCA half-site shared by most of the nuclear receptor superfamily, so all the family-level information lives in the two or three A-T bases of the 5' extension, and there is no spacer to check. One or two substitutions convert a predicted NBRE into an SF-1 element or into one half of a genuine RXR heterodimer site. Prediction of nuclear receptor elements improves substantially when dimer configuration is modeled, which is unavailable here (Sandelin and Wasserman 2005) 15563547. Treat a hit as "a nuclear receptor half-site" first. Provenance is comparatively good here: NR4A1.H14CORE.0.PSM.A is grade A with ChIP, HT-SELEX and methyl-HT-SELEX support and NR4A2.H14CORE.0.SM.B has both in vitro assays, so AAAGGTCA is well supported as directly bound. The one NR4A3 model, CIS-BP M00818_2.00, is inferred from the mouse protein at 97% DNA-binding-domain identity rather than measured for human NR4A3 (Weirauch et al. 2014) 25215497.
How this was worked out
The NGFI-B response element was defined as AAAGGTCA by genetic selection in yeast, showing that the receptor recognizes a single half-site carried on an A/T-rich 5' extension rather than a spaced repeat.
Still open then: Why three extra base pairs 5' of the half-site were needed was not yet explained structurally.
Domain swaps between NGFI-B and SF-1 established monomer binding as a third paradigm of nuclear receptor-DNA interaction alongside homodimers and RXR heterodimers, and traced the different 5' requirements of the two receptors to residues in the A box.
Still open then: The minor-groove contacts inferred from swaps had no direct structural support.
NGFI-B and NURR1 were shown to heterodimerize with RXR on a direct repeat and to respond to retinoid signaling, so the same receptors read at least two unrelated DNA arrangements and a monomeric half-site is not the whole story.
Nur77-null mice were found to have unimpaired thymic and peripheral T cell death, overturning the expectation from dominant-negative experiments that the receptor is required for negative selection and pointing instead to redundancy among the three paralogues.
A dimeric arrangement, the NurRE, was shown to mediate CRH-induced POMC transcription in pituitary cells, establishing an everted pair of NBRE-related octamers as the configuration that carries a physiological signal.
The crystal structure of the NGFI-B DNA-binding region on the NBRE showed a single receptor molecule using its zinc modules on the AGGTCA half-site and an A-box substructure reaching into the minor groove of the 5' A-T base pairs, the first structure of a nuclear receptor bound to DNA as a monomer.
The NURR1 ligand-binding domain was found to have no accessible ligand pocket, its interior filled with bulky side chains, defining a class of receptors that activate without a ligand.
Still open then: Whether the domain can still be targeted pharmacologically.
Deleting all three NR4A genes in CAR T cells relieved an exhaustion-like transcriptional program and improved tumor control, demonstrating at the level of the whole subfamily the redundancy that single knockouts had obscured.
Still unresolved
- Whether an isolated NBRE is functional in vivo is unsettled: the monomeric element is well measured in vitro, yet the configurations shown to respond to physiological signals are the dimeric NurRE and the NBRE-RXR direct repeat, and a lone octamer is often weak in reporter assays.
- Whether the three paralogues have distinct target sets cannot currently be answered: their motifs are identical, single knockouts are near-normal, and the clearest phenotypes appear only when all three are removed together.
- Whether the ligand-binding domain can be engaged pharmacologically despite the absence of a conventional pocket remains contested, with the crystal structure showing the interior filled by side chains while small molecules continue to be reported as acting through that domain.
- The boundary between this family and the RXR heterodimers is not settled in the data: one Vierstra archetype assigned here is a bare AGGTCA half-site flagged as belonging with the RXR heterodimers, and the JASPAR NR4A2::RXRA entry is a dimer model whose nearest neighbors lie outside the family.
The sequence identifies a nuclear receptor half-site, not a receptor. AAAGGTCA constrains eight positions, six of which are the AGGTCA half-site used across the superfamily, so tens of thousands of chance matches exist and one or two substitutions turn a predicted NBRE into an SF-1 element or into half of an RXR heterodimer site. Nothing separates NR4A1, NR4A2 and NR4A3. The AAAGGTCA string itself is well supported in vitro (a grade-A HOCOMOCO model with HT-SELEX and methyl-HT-SELEX evidence), but the only NR4A3 model is inferred from mouse, one Vierstra archetype in this family is flagged as probably belonging with the RXR heterodimers, and the JASPAR NR4A2::RXRA entry is a dimer model rather than an NR4A monomer motif. Occupancy requires the receptor to have been induced.
References
- Wilson TE, Fahrner TJ, Johnston M, et al. Identification of the DNA binding site for NGFI-B by genetic selection in yeast. Science 1991;252:1296-300. PMID 1925541
- Wilson TE, Fahrner TJ, Milbrandt J. The orphan receptors NGFI-B and steroidogenic factor 1 establish monomer binding as a third paradigm of nuclear receptor-DNA interaction. Mol Cell Biol 1993;13:5794-804. PMID 8395013
- Perlmann T, Jansson L. A novel pathway for vitamin A signaling mediated by RXR heterodimerization with NGFI-B and NURR1. Genes Dev 1995;9:769-82. PMID 7705655
- Labelle Y, Zucman J, Stenman G, et al. Oncogenic conversion of a novel orphan nuclear receptor by chromosome translocation. Hum Mol Genet 1995;4:2219-26. PMID 8634690
- Lee SL, Wesselschmidt RL, Linette GP, et al. Unimpaired thymic and peripheral T cell death in mice lacking the nuclear receptor NGFI-B (Nur77). Science 1995;269:532-5. PMID 7624775
- Philips A, Lesage S, Gingras R, et al. Novel dimeric Nur77 signaling mechanism in endocrine and lymphoid cells. Mol Cell Biol 1997;17:5946-51. PMID 9315652
- Zetterström RH, Solomin L, Jansson L, et al. Dopamine neuron agenesis in Nurr1-deficient mice. Science 1997;276:248-50. PMID 9092472
- Meinke G, Sigler PB. DNA-binding mechanism of the monomeric orphan nuclear receptor NGFI-B. Nat Struct Biol 1999;6:471-7. PMID 10331876
- Wang Z, Benoit G, Liu J, et al. Structure and function of Nurr1 identifies a class of ligand-independent nuclear receptors. Nature 2003;423:555-60. PMID 12774125
- Sandelin A, Wasserman WW. Prediction of nuclear hormone receptor response elements. Mol Endocrinol 2005;19:595-606. PMID 15563547
- Weirauch MT, Yang A, Albu M, et al. Determination and inference of eukaryotic transcription factor sequence specificity. Cell 2014;158:1431-1443. PMID 25215497
- Chen J, López-Moyado IF, Seo H, et al. NR4A transcription factors limit CAR T cell function in solid tumours. Nature 2019;567:530-534. PMID 30814732
10 motifs in this family — JASPAR 3 · HOCOMOCO 2 · CIS-BP 3 · Vierstra 2
| motif | database | consensus | bits |
|---|---|---|---|
| AC0119:NR4A:Nuclear_receptor | Vierstra | GACCTTTAAAGGTCA | 25.1 |
| NR4A2::RXRA (MA1147.2)dimer | JASPAR | aGGTCaTTGACCc | 14 |
| NR4A1.H14CORE.0.PSM.APSMA | HOCOMOCO | AAAGGTCA | 13.1 |
| M08153_2.00 | CIS-BP | AAAGGTCA | 12.6 |
| NR4A1 (MA1112.3) | JASPAR | AAAGGTCA | 12 |
| NR4A2.H14CORE.0.SM.BSMB | HOCOMOCO | AAAGGTCA | 11.7 |
| NR4A2 (MA0160.3) | JASPAR | AAAGGTCA | 11.5 |
| M05639_2.00 | CIS-BP | AAAGGTCA | 10.7 |
| M00818_2.00Mus musculus 0.97 | CIS-BP | AAAGGTCA | 9.7 |
| AC0114:NR4A/NR2C:Nuclear_receptorlooks like RXR_HETERO | Vierstra | aAGGTCA | 9 |
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.