NRF1

The factors

This family has exactly one human member: NRF1, nuclear respiratory factor 1, also called alpha-PAL for "alpha palindromic-binding protein". It was defined functionally, by mutational and binding analysis of the somatic cytochrome c promoter, before it was cloned ^2166701. Its DNA-binding domain lies in the N-terminal half of the protein and belongs to no previously described class; the closest relatives are the sea urchin developmental regulator P3A2 and the Drosophila protein Erect wing (EWG), and the three define a small, conserved family unrelated to zinc fingers, helix-turn-helix, bZIP or any other common fold ^8253388. The C-terminal half, which carries the activation function, is divergent between the three ^8253388. NRF1 binds as a homodimer ^38055835 and is broadly expressed; it is a housekeeping factor rather than a tissue-specific one, and its activity is modulated by phosphorylation of an N-terminal serine cluster and by coactivation from the PGC-1 family ^18391175. Because there is one gene, the usual paralogue problem does not arise — but a nomenclature problem does, described under Reading a hit.

The motif

NRF1 recognizes a palindrome, GCGCATGCGC, and the palindromic character is the defining feature: each subunit of the homodimer contacts one GCGC half, and the guanine contacts on the two strands mirror each other ^{2166701,38055835}. The motif models in this collection extend it slightly, to an 11- or 12-mer — the HOCOMOCO and Vierstra consensus is TGCGCATGCGC — so the element is equally well described as a tandem pair of TGCGC half-sites separated by a single base ^38055835. The palindrome is the older rendering; the extended repeat is what the matrices encode. Some models (JASPAR MA0506.3, CIS-BP M08177_2.00) place a C rather than an A/T at the central position, giving TGCGCCTGCGC, which reflects tolerance at the one position that is not part of either half-site ^38055835.

Two features follow directly from the composition and matter in practice. First, the site contains two CpG dinucleotides, and both fall in the positions that carry the most information. Second, it is therefore methylation-sensitive: cytosine methylation within the site blocks NRF1 binding, and in cells lacking DNA methyltransferases NRF1 gains occupancy at sites it cannot bind in wild-type cells. NRF1 was the factor used to demonstrate this competition genome-wide ^26675734, and it also scores as methylation-avoiding in systematic methyl-SELEX surveys ^28473536. Family cohesion is 0.94, the highest tier in this atlas: all four models are effectively the same matrix.

Where it acts

NRF1 sites are promoter elements, not distal enhancers, and they are strongly associated with CpG islands — which follows from the motif itself, since an unmethylated CpG-rich 10-mer is exactly what a CpG island is made of. The original targets were nuclear genes for mitochondrial proteins: somatic cytochrome c, cytochrome c oxidase subunits, a ubiquinone-binding subunit of the reductase complex, and the RNA component of RNase MRP, which functions in mitochondrial DNA replication ^2166701. NRF1 also activates the promoter of TFAM, the mitochondrial transcription factor A gene, which is the direct mechanistic link between a nuclear promoter element and mitochondrial genome expression ^8108407. Beyond respiratory genes, NRF1 sites occur in many broadly expressed promoters, frequently alongside GABP (historically NRF-2), SP1 and YY1 sites ^{8383622,18391175}. An exact 10-mer of this composition is expected only a few thousand times in 3 Gb of uniform sequence, and because CpGs are depleted outside CpG islands, genuine matches concentrate in island promoters rather than being spread across the genome.

Biology and disease

NRF1 is the central nuclear factor coordinating expression of the respiratory apparatus with the mitochondrial genome. Mouse Nrf1 knockouts die at peri-implantation stages with depleted and unstable mitochondrial DNA ^11134350, a phenotype consistent with loss of Tfam and related targets rather than with a single metabolic defect. Through TFAM, TFB1M/TFB2M and the mitochondrial ribosome and import machinery, NRF1 sets the capacity of oxidative phosphorylation, and its activity is the point at which PGC-1-family coactivators act during mitochondrial biogenesis in muscle and brown fat ^18391175. The methylation dependence also gives NRF1 a role as a readout of the epigenome: in promoters that become hypermethylated, NRF1 occupancy is lost even though the sequence is intact ^26675734, so changes in NRF1 target expression can reflect methylation change rather than any change in the factor.

Reading a hit

The most common error with this family is not motif-related at all, it is naming. "NRF1" in this atlas means nuclear respiratory factor 1 and nothing else. It is not NFE2L1, which is also routinely abbreviated Nrf1 and belongs to the CNC-bZIP antioxidant-response family; it is not NFE2L2/NRF2; and it is not NRF-2, which turned out to be GABP, an ETS-domain factor ^8383622. A hit on GCGCATGCGC says nothing about antioxidant response elements or the KEAP1-NRF2 pathway.

With one gene in the family, a match is unusually close to a protein-level call, and the evidence supports direct binding: HOCOMOCO's NRF1.H14CORE.0.PS.A combines ChIP-seq with HT-SELEX at grade A, so this is not a ChIP-only inference. The 10-11-mer is long and information-rich enough that a strong match in a CpG-island promoter is genuinely suggestive. What a match does not tell you is occupancy, and here the gap is specific and checkable: because the site carries two CpGs and NRF1 will not bind a methylated site, a perfect match inside a methylated region is probably unbound, so check methylation state ^26675734. Finally, the Vierstra archetype here is AC0555:NRF:Unknown: the class is recorded as unknown, and the label names clustered factor annotations, not a validated protein.