Rerun the 2002 SPR with the two controls it did not have: cleave the GST, and put the protein on the chip instead of the DNA
also called the missing SPR controls · orientation swap
The 33 nM measurement used a two-headed protein flowed over a surface densely coated with DNA, which is the standard way to get an artificially tight number. Two straightforward repeats - remove the second head, and swap which partner is stuck down - would show how much of the number is real.
The same thing, for a scientist
Birke's SPR reports KD 3.28 +/- 1.7 x 10^-8 M for GST-MT over immobilised biotinylated CpG52. Four features of that geometry inflate apparent affinity and none is controlled: an obligate GST homodimer presenting two MT domains; DNA as the immobilised ligand at the density produced by two 5-pmol injections onto a streptavidin chip; a 52-bp duplex with six CpG sites over which the paper's own DNase I footprinting shows three separate protected regions; and a kinetic KD taken as k_d/k_a from initial rates, the form most inflated by rebinding on a multivalent surface. Three repeats decide how much is real: cleave the GST and rerun; reverse the orientation so the protein is captured and DNA flows; and titrate immobilisation density. Each is a day on an instrument most structural laboratories already own. The reported SD is 52% of the mean, so the original measurement is also less precise than its two significant figures imply.
The three repeats
1. Cleave the tag. Express GST-MT from the same construct, cleave with Factor Xa (pGEX-3X's protease) or subclone into a TEV- or SUMO-cleavable vector, confirm monomeric by SEC-MALS, and repeat the identical SPR. This is the single control the original lacks, and Erfurth's published explanation of the whole discrepancy depends on its outcome. A prediction: if GST bivalency were the dominant term, the tagless protein should lose most of the 19-fold gap to Risner's GST-tagged short construct - but ana-cxxc-dna-affinity-ladder already bounds GST-plus-salt at about 5.5-fold on a matched pair, so the expected loss is modest and the rest must come from elsewhere.
2. Reverse the orientation. Capture the protein on the chip - via an anti-GST surface or a His-capture surface - and flow the DNA duplex as analyte. A monovalent analyte over a surface removes the avidity contribution of multiple immobilised DNA molecules in one step. If the affinity falls by an order of magnitude on orientation swap alone, the 33 nM is an avidity number and should be recorded as such.
3. Titrate the surface density. Immobilise at a tenth and a hundredth of the original loading and refit. An avidity artefact is density-dependent and a true monovalent affinity is not; this is the cheapest of the three and it is diagnostic on its own. Birke injected 5 pmol twice, which is a high loading for a 52-bp duplex on a standard SA chip.
And refit as an equilibrium isotherm rather than k_d/k_a. Steady-state fitting at each concentration gives a KD that does not inherit the rebinding artefact in k_d, and a discrepancy between the kinetic and steady-state values is itself the diagnostic.
| design | Three SPR repeats of Birke's measurement: (1) tagless MT domain after protease cleavage, SEC-MALS-confirmed monomeric, identical chip and buffer; (2) orientation reversed - protein captured, DNA flowed as monovalent analyte; (3) immobilisation density titrated over two orders of magnitude. All three refit as steady-state isotherms as well as k_d/k_a. |
|---|---|
| readout | KD by kinetic fit and by steady-state fit, at each tag state, orientation and surface density |
| controls | SEC-MALS oligomeric state on every preparation; methylated CpG52 as the specificity negative (Birke reports greater than 1 mM); a reference flow cell with GpC52; the original GST-tagged high-density condition rerun in the same session as the internal benchmark |
| cost tier | low |
| timeline | three days on the instrument once protein is in hand; cleavable construct is the rate-limiting step |
| what it closes | whether the 33 nM figure is a monovalent affinity, a bivalent avidity, or a surface artefact - which determines whether the tight end of the affinity ladder is a real property of the fusion's DNA contact |
| shown for fusion | NO - isolated-domain biophysics by design. |
predictions
[ "Density titration will show a density-dependent KD if avidity is operating, and a density-independent one if not. This is diagnostic on its own.", "Orientation swap will weaken the apparent affinity if surface avidity is the main term.", "Tag cleavage will cost less than an order of magnitude, because the GST-plus-salt term is already bounded at about 5.5-fold by a matched solution pair.", "Steady-state and kinetic fits will diverge if rebinding is inflating k_d." ]
- If the 33 nM survives every control, the CXXC-DNA contact is comparable to or tighter than the menin contact and the potency bar for a DNA-competitive ligand is nanomolar - which makes the CXXC target substantially less attractive than the phase-2 dossier currently assumes.
- If the 33 nM collapses to micromolar under the controls, the graph's original argument is restored but on a number nobody has yet measured at physiological salt, so the entries would have to be rewritten a second time.
- Birke's orientation - DNA immobilised, protein in flow - may be the physiologically apt one, in which case 'controlling out' the avidity removes exactly the feature that makes the measurement relevant to chromatin.
- Should the repeat use the original BIAcore chemistry or a modern instrument? A modern BLI or SPR instrument changes the surface chemistry and would confound a direct comparison with the 2002 numbers; running both is the conservative choice.
- Is the relevant biological geometry actually the DNA-immobilised one? Chromatin is the immobilised partner in a nucleus, so Birke's orientation may be the more physiological of the two, and the orientation swap tests the artefact rather than the biology.
- Could a nucleosome be used as the immobilised ligand instead of naked duplex? That would be the first measurement of this contact on a physiological substrate and nobody has attempted it.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Birke's SPR immobilised the DNA on a streptavidin chip by two injections of 5 pmol of biotinylated duplex, flowed GST-MT as analyte, and computed KD as the ratio of dissociation to association rate from initial rates, giving 3.28 +/- 1.7 x 10^-8 M over eight determinations; the paper includes no tagless control, no orientation swap and no density titration, and its own DNase I footprinting shows three separate protected regions on one selected fragment.
The MT domain of the proto-oncoprotein MLL binds to CpG-containing DNA and discriminates against methylation. · Birke M, Schreiner S, García-Cuéllar MP, et al. · Nucleic Acids Res 2002What this evidence cannot show
Read from the paper's Methods and Results this cycle via the PMC HTML route. Judging a 2002 BIAcoreX experiment by 2026 SPR practice is anachronistic - density titration and orientation swap were not standard then - and the criticism is of how the number is used now, not of the work. The footprinting was on a different, selected 74-bp fragment rather than on CpG52 itself, so 'three binding sites per duplex' is an inference about the ligand class, not a direct observation on the SPR ligand.
Evidence that argues against this (1)
- Multivalency does not always inflate a measured KD: if the analyte is bivalent and the surface is sparse, a 1:1 fit can be accurate. The diagnostic is the density titration, which is why it is the cheapest and most decisive of the three repeats, and until it is run the avidity attribution is an expectation rather than a finding. Erfurth FE et al. 2008
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Erfurth et al. attribute the roughly two-orders-of-magnitude difference between the SPR figure and their ITC figures specifically to GST dimerisation and to the potential for bivalent interaction of GST-MT with two sites on the DNA, citing prior work that GST dimerisation influences fused DNA-binding domains.
MLL protects CpG clusters from methylation within the Hoxa9 gene, maintaining transcript expression. · Erfurth FE, Popovic R, Grembecka J, et al. · Proc Natl Acad Sci U S A 2008What this evidence cannot show
An explanation offered in a discussion paragraph, supported by a citation to general GST behaviour rather than by any experiment on this protein. Eighteen years later no one has tested it, which is the reason this experiment entry exists.
Evidence that argues against this (1)
- The GST-plus-salt term is bounded at about 5.5-fold by Erfurth's own CpG1-I ITC against Risner's FP on the same oligonucleotide and the same construct, so GST dimerisation cannot be the whole of a 1,300-fold gap and is probably not the whole of the 19-fold residual either. The explanation this experiment tests is already known to be insufficient; the experiment establishes by how much. Risner LE et al. 2013
The SPR controls price the alternative explanations for the same 19-fold step.
Determines whether the 33 nM end of the ladder is a monovalent affinity, an avidity, or a surface artefact.
Re-measures the KMT2A MT domain-DNA contact with the controls the 2002 experiment lacked.