Attaching a peptide to a surface looks like a reagent choice and is really a design decision. Get the orientation wrong and a fraction of the material is inert. Omit the spacer and the binding partner cannot reach. Put too much on the surface and the constant that comes out is distorted by rebinding and transport rather than reporting affinity. This page covers the three decisions that determine whether an immobilised peptide reports what you think it reports, whether the surface is a sensor chip, a bead, or a plate.
A peptide coupled through its lysines will attach through whichever lysine reacted first, and different molecules will attach through different ones. The surface then carries a mixture of orientations, and every molecule whose binding determinant happens to face the support is dead weight — present in the loading calculation, absent from the result. On a sequence with several lysines this is not a small correction.
Site-specific attachment solves it. A cysteine added at one terminus gives a single unambiguous attachment point through maleimide or haloacetyl chemistry; a biotin placed at one terminus gives the same thing through streptavidin capture. Which terminus is the real question, and the rule is simple: attach the end that is not doing the binding. If an alanine scan or a truncation series has shown the activity lives in the C-terminal half, the handle goes on the N-terminus.
A peptide attached directly to a support sits against it. The support is not a neutral background: it occupies space, it carries its own charge, and it sterically blocks approach from one whole hemisphere. For a short peptide the binding site may be only a few ångströms from the attachment point, which means it is effectively buried.
A spacer — commonly a short PEG unit, an aminohexanoic acid, or a few glycines and serines — lifts the peptide clear. The usual failure is omitting it entirely rather than choosing the wrong length. A sensible default is enough length to clear the surface layer, with the recognition that an unnecessarily long flexible spacer costs entropy and can allow the peptide to fold back onto the support. Where the peptide is short and the attachment point close to the active region, the spacer is not optional.
This is the one that silently corrupts data. Two separate problems arise from a crowded surface.
Rebinding and avidity. When a partner dissociates from a densely packed surface, it does not escape into bulk solution; it encounters a neighbouring peptide and binds again before diffusing away. The measured dissociation rate is slowed by geometry rather than by chemistry, and since affinity is derived from that rate the result is an affinity that looks far better than it is. The same effect applies with force when the partner itself has more than one binding site: a multivalent analyte can engage several immobilised molecules at once, or move along the surface without ever fully letting go, producing a gross overestimate of affinity1. Because the distortion acts through the dissociation rate, it is also the hardest kind to spot by eye: the sensorgram looks clean, the fit looks good, and the answer is wrong.
Mass transport limitation. If the surface binds analyte faster than diffusion can deliver it, the measured association rate reports diffusion rather than chemistry. The principal fix is to reduce the surface density of immobilised ligand, which is why the standing advice is to use the lowest density that still gives an adequate signal2. Raising the flow rate and fitting with a model that accounts for transport address the same problem, and in practice the three are used together. In practice the target is lower than newcomers expect: a common rule of thumb for kinetic work is to choose a ligand level giving a maximum analyte response of roughly 10–30 response units, rather than the several hundred that is perfectly adequate for a yes/no binding check4. Densities above a few hundred response units are a recognised risk factor for transport-limited kinetics2. Where avidity is a specific concern the immobilisation level may be reduced further still, with a protein blocking step added after coupling3.
The diagnostic test costs one extra surface. Prepare the same peptide at two clearly different densities and measure the kinetics on both. If the fitted constants agree, density is not distorting them. If the apparent association rate falls as density rises, or the dissociation rate slows, the surface is interfering and the lower-density result is the one to trust2. Varying the flow rate is the companion test: a genuine chemical rate does not care how fast the buffer moves past.
In a binding assay the peptide stays put. In a pull-down it has to come off again, with its catch, and that imposes additional design constraints.
An immobilised-peptide experiment needs a control surface prepared identically in every respect except the sequence: same chemistry, same spacer, same density, scrambled or point-mutated sequence. A blank surface is not sufficient, because it controls for the support but not for the peptide's generic properties — charge, hydrophobicity, and the simple fact of a peptide being there. Where the peptide carries a known binding motif, a single-residue mutant that abolishes the motif is the strongest control available. See control peptides in experimental design for the general principles and designing peptides for binding studies for sequence-level choices.
Finally, remember what an immobilised measurement is. Tethering the peptide changes its entropy, its accessible conformations and its effective local concentration, so a surface-derived constant is not interchangeable with one measured in free solution — see peptide versus protein binding affinity and interpreting peptide binding data.
| Orientation | Site-specific via a terminal Cys or biotin; attach the non-binding end |
| Spacer | Usually essential; PEG, Ahx or a short Gly/Ser run |
| Density | Lowest that gives adequate signal |
| Density artefacts | Rebinding and avidity slow the apparent off-rate; mass transport distorts the on-rate |
| Test | Two densities, and two flow rates — do the constants agree? |
| Control | Scrambled or point-mutant peptide at matched density, not a blank surface |
LifeTein supplies peptides made for immobilisation: terminal cysteine for thiol coupling, N- or C-terminal biotin with or without a cleavable linker, defined PEG and aminohexanoic acid spacers, and matched scrambled or point-mutant control sequences made in the same run. Tell us the capture chemistry and which end does the binding, and we will put the handle on the right terminus.