PURA / PURB
PURA and PURB (Pur-alpha and Pur-beta) are PUR-repeat proteins that bind purine-rich DNA and RNA, with a strong preference for single strands rather than duplex. The PUR element was defined as a repeated purine-rich sequence upstream of the human c-MYC P1 promoter, at a site of DNA bending inside a replication initiation zone, and was given the consensus GGNNGAGGGAGARRRR. The PUR repeats fold into a Whirly-like domain, and Pur-alpha can unwind double-stranded DNA rather than simply read it. These proteins act mostly as repressors, best characterized on the smooth muscle alpha-actin promoter, where they bind the GGN-repeat strand of core enhancer elements. De novo PURA variants cause PURA syndrome, a neurodevelopmental disorder with neonatal hypotonia, seizures and severe developmental delay.
The factors
Three PUR-repeat genes exist in humans: PURA (Pur-alpha), PURB (Pur-beta) and the less studied PURG. PURA and PURB are broadly expressed, with PURA especially abundant in brain. The proteins are built from tandem PUR repeats; crystallography of Pur-alpha showed that two repeats pack together into a single Whirly-like domain, a fold otherwise associated with plant single-stranded nucleic-acid-binding proteins, presenting an unusual continuous binding surface rather than the discrete base-reading elements of a zinc finger or helix-turn-helix 19846792. A third repeat mediates dimerization between molecules 19846792. This architecture explains the central functional fact about the family: PUR proteins are not conventional double-stranded sequence-specific transcription factors. They bind purine-rich single-stranded DNA and RNA with high affinity, and Pur-alpha can actively unwind duplex DNA rather than reading it intact 26744780,23018800. Pur-alpha also binds RNA, including trinucleotide-repeat transcripts 26744780, and localizes to stress granules, so its functions extend beyond transcription into RNA handling and, in the original work, into DNA replication 1545807.
The motif
The literature element comes from mapping a site of DNA bending 1.6 kb upstream of the human c-MYC P1 start site, near the center of a reported replication initiation zone. Methylation interference mapped specific guanosine contacts in a 24-mer, and the derived consensus was GGNNGAGGGAGARRRR, a sixteen-base purine-rich element found at replication origins and in gene flanking regions from yeast to human 1545807. In practice the element is usually described more loosely as a GGN repeat, and binding is strand-selective: the purine-rich strand is bound, the complementary pyrimidine-rich strand is not 1545807,16436378. There is no half-site grammar, no spacing rule, and no palindrome. The data here disagree with this, and the disagreement is worth stating plainly. This family contains exactly one motif, the HOCOMOCO model PURB.H14CORE.0.B.B, whose consensus is ACCCAC, six bases and only 7.6 bits of information. Its reverse complement, GTGGGT, is GT-rich rather than purine-rich in the GGN sense, it is far shorter than the published PUR element, its evidence code B marks it as a protein-binding-microarray (PBM) measurement rather than derived from HOCOMOCO's own ChIP-seq or SELEX data, and its nearest neighbours across all four databases lie outside this family (best cross-family similarity 0.73, label agreement 0). There is no PURA motif in these databases at all, and with a single member the family has no internal cohesion to measure. The honest position is that the published PUR element is a single-strand, purine-rich sequence that a position-weight matrix over double-stranded DNA is the wrong tool for, and that the one matrix available under this key should not be treated as a validated model of PUR binding.
Where it acts
Documented PUR-protein sites are promoter-proximal and, in the original work, at replication origins. The best-characterized regulatory target is the mouse and human smooth muscle alpha-actin promoter, where Pur-alpha and Pur-beta bind the GGN-repeat-containing strands of several cis elements within a core enhancer and repress transcription 16436378. Knockdown showed Pur-beta to be the dominant repressor in fibroblasts, with the combined action of both proteins needed for full repression, and the two proteins had indistinguishable binding affinity and specificity in vitro but distinct protein-interaction profiles, which is how cell-type-specific behavior arises from a shared site 16436378. PUR proteins also occur at the c-MYC upstream region where the element was found, and Pur-alpha associates with RNA and with cytoplasmic granules in neurons 19846792,23018800, so not all of its binding is chromatin-associated.
Biology and disease
Pura-null mice are viable at birth but develop severe neurological abnormalities with arrested postnatal brain development and impaired proliferation of neuronal and hematopoietic precursors 19846792,23018800. In humans, de novo PURA variants and 5q31.3 microdeletions that remove PURA cause PURA syndrome: profound neonatal hypotonia, feeding difficulty, seizures or seizure-like movements, and severe developmental delay with little or no speech 25342064,25439098,29097605. An early series proposed that disruption of PUR repeat III gives a more severe phenotype 25342064, and the Pur-alpha-DNA structure shows how some missense changes would impair nucleic-acid binding 26744780, but a study of 32 individuals found no significant correlation between mutation class and severity 29097605. PURB has no established monogenic disease association.
Reading a hit
Do not read a hit under this key as evidence of PUR binding. The only matrix available is a six-base, 7.6-bit model that does not resemble the published purine-rich PUR element and whose closest relatives in the motif databases belong to other families. More fundamentally, PUR proteins prefer single-stranded purine-rich DNA and can unwind duplex 26744780, so scanning double-stranded genomic sequence with a PWM is not a test of PUR binding even in principle. If you care about PUR proteins at a locus, the useful evidence is a long GGN-rich or polypurine tract on one strand, ideally at a site with known single-strand character, plus direct biochemical or crosslinking data. Treat any automated PURB annotation in a scan result as a placeholder.
How this was worked out
A HeLa protein of about 28 kDa binds the purine-rich strand of a repeated element (PUR) at a DNA bend 1.6 kb upstream of c-MYC, near a replication initiation zone; methylation interference maps guanosine contacts and gives the consensus GGNNGAGGGAGARRRR, also found at other origins and gene flanks from yeast to human.
Still open then: The identity of the protein and whether single-strand binding operates on chromosomal DNA.
At the smooth muscle alpha-actin promoter, purified Pur-alpha and Pur-beta bind GGN-repeat strands of several enhancer elements with essentially identical affinity and specificity, and repress transcription; cell-type differences come from their distinct protein partners, with Pur-beta the dominant repressor in fibroblasts.
Crystal structure of the Pur-alpha core: two PUR repeats pack into a Whirly-like PUR domain that binds nucleic acid through concave beta-sheet surfaces, and a third repeat mediates dimerization.
De novo PURA mutations are found in children with severe neurodevelopmental delay, hypotonia and seizures, and PURA is identified as the gene behind the neurological core of 5q31.3 microdeletion syndrome.
Still open then: Whether mutation position predicts severity; repeat III disruption was proposed to be worse.
Crystal structure of the Pur-alpha DNA/RNA-binding domain on single-stranded DNA shows base-specific recognition, DNA and RNA bound the same way, and the mechanism of duplex unwinding.
Across 32 new and 22 published individuals with PURA syndrome, no significant correlation is found between mutation class and severity, against the earlier suggestion that repeat III disruption is more severe.
Still unresolved
- How much sequence PUR proteins actually read is unsettled: the original mapping gave a sixteen-base consensus (GGNNGAGGGAGARRRR), later functional work describes the sites only as GGN repeats, and the Pur-alpha-ssDNA structure shows base-specific contacts without by itself defining a consensus.
- The only matrix under this key, the HOCOMOCO PURB model ACCCAC (GTGGGT on the other strand), does not resemble the published purine-rich element, and whether PURB has a genuine double-stranded preference that this model captures, or the model reflects weak signal, is untested.
- Whether PURA mutation type predicts severity is disputed: an early series proposed that disruption of PUR repeat III is more severe, whereas a study of 32 individuals found no significant correlation between mutation class and severity.
This is the weakest family in the atlas. It contains one motif, a six-base 7.6-bit HOCOMOCO model for PURB from a protein-binding microarray, with no PURA motif and no internal cohesion. That model (ACCCAC) does not match the published purine-rich PUR element, GGNNGAGGGAGARRRR, and its nearest neighbours in all four databases lie in other families. PUR proteins bind single-stranded purine-rich DNA and RNA and can unwind duplex DNA, so a double-stranded PWM scan is not a valid test of their binding. A hit here should be treated as an annotation placeholder, not as evidence that PURA or PURB binds the sequence.
References
- Bergemann AD, Johnson EM. The HeLa Pur factor binds single-stranded DNA at a specific element conserved in gene flanking regions and origins of DNA replication. Mol Cell Biol 1992;12:1257-65. PMID 1545807
- Graebsch A, Roche S, Niessing D. X-ray structure of Pur-alpha reveals a Whirly-like fold and an unusual nucleic-acid binding surface. Proc Natl Acad Sci U S A 2009;106:18521-6. PMID 19846792
- Weber J, Bao H, Hartlmüller C, et al. Structural basis of nucleic-acid recognition and double-strand unwinding by the essential neuronal protein Pur-alpha. Elife 2016;5. PMID 26744780
- Knapp AM, Ramsey JE, Wang SX, et al. Nucleoprotein interactions governing cell type-dependent repression of the mouse smooth muscle alpha-actin promoter by single-stranded DNA-binding proteins Pur alpha and Pur beta. J Biol Chem 2006;281:7907-18. PMID 16436378
- Johnson EM, Daniel DC, Gordon J. The pur protein family: genetic and structural features in development and disease. J Cell Physiol 2013;228:930-7. PMID 23018800
- Hunt D, Leventer RJ, Simons C, et al. Whole exome sequencing in family trios reveals de novo mutations in PURA as a cause of severe neurodevelopmental delay and learning disability. J Med Genet 2014;51:806-13. PMID 25342064
- Lalani SR, Zhang J, Schaaf CP, et al. Mutations in PURA cause profound neonatal hypotonia, seizures, and encephalopathy in 5q31.3 microdeletion syndrome. Am J Hum Genet 2014;95:579-83. PMID 25439098
- Reijnders MRF, Janowski R, Alvi M, et al. PURA syndrome: clinical delineation and genotype-phenotype study in 32 individuals with review of published literature. J Med Genet 2018;55:104-113. PMID 29097605
- Vorontsov IE, Eliseeva IA, Zinkevich A, et al. HOCOMOCO in 2024: a rebuild of the curated collection of binding models for human and mouse transcription factors. Nucleic Acids Res 2024;52:D154-D163. PMID 37971293
1 motifs in this family — HOCOMOCO 1
| motif | database | consensus | bits |
|---|---|---|---|
| PURB.H14CORE.0.B.BBB | HOCOMOCO | aCCcAC | 7.6 |
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.