In this paper regarding a polymorphism within the AQP7 gene and its potential effects on one's metabolic profile, the authors use a luciferase assay to confirm that the alleles of this SNP have a definite impact on the expression of AQP7 in the presence of CEBPB, but the SNP itself falls outside the predicted CEBPB binding sequence: enter image description here

The only thing I can imagine is a second transcription factor that cooperates with CEBPB has a binding sequence that overlaps this SNP, but I see no mention of such a thing in this paper. If this SNP truly affects the binding of CEBPB, then wouldn't the corresponding binding sequence have to extend through this SNP?


To quote the paper, "A gel shift assay was then performed to investigate whether the A-953G SNP affects C/EBP binding. Compared with the radiolabeled oligonucleotide carrying the 953A, the one carrying the -953G variant showed a 30+/-7% reduced binding to C/EBP"

And here is the gel shift assay:

enter image description here

As far as I can tell, even if the scientists had reason to believe that this SNP affected CEBP binding in this region, it shouldn't be beyond science for someone to hypothesize how that is.

  • 1
    $\begingroup$ The authors of the paper presumably do not say — I assume you have read the discussion and this is the case — because they have no experimental evidence that bears on the question. They are professional scientists and are behaving as such. We can all think of possibilities, but it is hardly fruitful to discuss them without evidence. $\endgroup$
    – David
    Oct 20 at 19:11
  • $\begingroup$ Did they assess binding of CEBPB to the two different promoter sequences? Have you looked up what other factors are known to bind this promoter? There isn't enough information here to give a non-speculative answer without doing a lot of extra research, but the short answer to your question at the end is 'no' — there are many ways that SNP could affect expression, only some of which involve altered binding of CEBPB. $\endgroup$
    – tyersome
    Oct 20 at 19:12
  • 1
    $\begingroup$ @David Well even if these scientists didn't wish to include any postulations in their publication concerning this, I still wish to know what can be said outside of the paper. When I said "I see no mention of such a thing in this paper" it wasn't to imply that the authors left something out or that they had an obligation to speak on it, my question stands regardless of whatever is or isn't in the paper I cited. $\endgroup$ Oct 20 at 20:12
  • 1
    $\begingroup$ If you have a question about the indirect influence of DNA regions on transcription factor binding, I think it would be better to pose in general terms, rather than relate it to a particular paper, which in any case few members of this list are likely to read. But you need to show evidence that you have thought about it yourself first. $\endgroup$
    – David
    Oct 20 at 20:23
  • 2
    $\begingroup$ I think the question is fine as-is. The question title asks about a general concept; the body of the question shows what motivated OP to ask the question and demonstrates that they've thought about it enough to rule out some obvious counter possibilities. I think it's fine that OP doesn't quite gather on their own the importance of structure in binding to DNA sequences, because a lot of general textbooks gloss over or fail to mention this when talking about recognized sequences, hence the question and resulting answer. $\endgroup$
    – Bryan Krause
    Oct 20 at 20:28

This is an answer to the general question posed in the title and not what is happening in the specific case of C/EBPβ.

How can a nucleotide falling outside a binding site affect a transcription factor's ability to bind to said site?

Transcription factors (TFs) associate with a specific shape of DNA. Two binding motifs that are identical at the sequence level may have different helical conformations due to torsional strain associated with flanking sequence context or proximal protein binding. See Figure 1 of Inukai et al. 2017 1 for an overview of the factors that affect TF-DNA interactions beyond primary DNA sequence:

Figure 1 from Inukai 2017 TF-level features: (A) Several TFs can display binding specificity for multiple, distinct nucleotide sequence motifs. Motifs shown are examples of two motifs bound by the bispecific human forkhead TF FOXN2: FHL (red box) and FkhP (blue box) binding site motifs; motifs were obtained from UniPROBE (Accession Number UP00521). Interactions between (B) TFs and (C) TFs and non-DNA-binding cofactors can specify distinct binding site motifs from the monomeric TF motif.
DNA-level features: (D) DNA modifications, such as 5-methylcytosine (left), can modulate TF binding. (E) Numerous TFs use DNA shape readout, such as minor groove width (depicted by red arrows), and rotational parameters such as helix twist, propeller twist, and roll, as part of TF–DNA recognition. (F) Sequences and features outside of the binding site motif (depicted by blue box), such as GC content and / or DNA shape, can modulate TF– DNA binding. These features may immediately flank the core binding site, or may extend more distally from the motif. (G) Genetic variation in either the TF protein sequence (depicted by orange star, middle) or the DNA binding site (depicted by X, right) can alter TF–DNA binding.


  1. Inukai S, Kock KH, Bulyk ML. Transcription factor-DNA binding: beyond binding site motifs. Curr Opin Genet Dev. 2017 Apr;43:110-119.
  • 1
    $\begingroup$ Thank you, situation B was essentially what I had imagined might be the case but this does broaden my horizons $\endgroup$ Oct 20 at 20:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.