Why peptide immunization gives you peptide binders
If you immunise with a peptide, most of what you can recover binds the peptide. Whether you ever reach the clones that work on the intact protein depends less on how hard you screen than on whether your platform can apply selective pressure at all.
Short answer: hybridoma and B-cell cloning fix the clone set at fusion and let you filter only afterwards, one clone at a time. Phage display applies the pressure before screening — you can pan on peptide, compete with the unmodified peptide, and pan on cells, so only clones that already survived every pressure reach the screening bench. For PTM targets, where protein-reactive clones are a small fraction of peptide-reactive ones, that difference decides whether the campaign returns anything usable.
The arithmetic that kills a screening-based campaign
A phospho-, methyl- or otherwise modification-specific campaign starts from a synthetic peptide, because that is the only practical way to present the modified site. The immune response is therefore dominated by antibodies that bind the peptide. That is not a flaw in the animal; it is what was presented.
The problem appears at the next step. Clones that bind the peptide and also detect the modification on the full-length protein in a lysate are a minority of that response. With hybridoma or B-cell cloning, the only way to find them is to screen: run every clone on peptide, then run the survivors on protein or on cells. Two assays over a panel that must be large enough to contain a rare event, and the panel size is fixed before you start.
Hybridoma and B-cell cloning sample on the order of 104–105 cells. If the clones you need are rarer than that, no amount of screening recovers them, and you find that out only after paying for the whole cascade. The probability of success was set at fusion; screening just reveals it, slowly and expensively.
Selection is not screening
Phage display changes where the filter sits. The library samples orders of magnitude more of the repertoire, and panning is itself the selective step rather than a readout. That makes the pressures composable:
- Pan on the modified peptide.
- Include an excess of soluble unmodified peptide during binding, so clones that cannot discriminate are competed away.
- Pan on cells or on the full-length antigen, so clones that only work on a synthetic fragment do not survive.
What reaches the screening bench has already passed all of it. The combinatorial screen that is impossible by hand becomes a short confirmation on a small number of clones. A practical consequence is that campaigns do not need to be scaled up with more animals or bigger fusions to raise the odds: the enrichment is done in the library, so a modest amount of starting material is enough.
The strategic difference for a sponsor is not cost per clone. It is that a screening-based campaign buys probability with budget, while a selection-based campaign buys it with campaign design.
Two published panels that show the failure mode directly
Both of the studies below were run on rabbit immunisation paired with Fab phage display, with Abwiz Bio co-authors and structures deposited in the PDB. In each, the tightest binder on the peptide was not the useful antibody.
Methylated lysine — the tightest clone failed on Western blot
| Clone | KD, trimethylated peptide | Detects full-length protein on Western blot? |
|---|---|---|
| E6 — tightest in panel | 0.90 nM | Faint band only |
| C9 | 14 nM | Yes — clean, SMYD3-dependent |
| D6 | 54 nM | Yes — clean, SMYD3-dependent |
| F9 | 300 nM | No — off-target bands only |
Co-crystal structures explained it. E6 and F9 hydrogen-bond the peptide’s free C-terminal carboxyl — a group that does not exist in a full-length protein, where the sequence simply continues. Amidating the peptide C-terminus nearly abolished binding for those two clones and left C9 and D6 unchanged. The tightest binder was tight for a reason that evaporates the moment you leave the peptide.
Ishii M et al. J Biol Chem 2021;296:100176. PMID 33303630. PDB 6LDV, 6LDW, 6LDX, 6LDY. Full write-up.
Phospho-Akt pSer473 — the tighter clone was the less selective one
| Clone | KD, phosphorylated peptide | KD, unmodified peptide | Selectivity |
|---|---|---|---|
| A4 | 3.34 nM | 1.12 µM | ~300-fold |
| C7 | 26.9 nM | No binding detected | Not quantifiable |
Binding the modified form tightly and ignoring the unmodified form are different properties, and ranking on the first does not select for the second. In this campaign an excess of soluble non-phosphorylated peptide was present during binding, which is what removed the non-discriminating clones before screening began.
Kasahara K et al. J Biol Chem 2024;300(12):107989. PMID 39542251. PDB 8JOW, 8ZPU, 8ZXW. Full write-up.
What to ask a CRO before a PTM campaign
- Where does the negative selection happen? If the answer is “we screen it out afterwards”, the panel size decides the outcome.
- Can the campaign pan on the real antigen, not only the peptide? Cells, membrane preparations or full-length protein, depending on the target.
- What is the deliverable screened against? If the antibody has to work on endogenous protein in a Western blot or an IHC section, that assay should decide the panel — not the peptide ELISA and not the affinity ranking.
- Is the output a panel or a single clone? Mechanistically distinct clones give options when the application changes; one clone gives none.
Frequently asked questions
Why does my phospho-specific antibody work on the peptide but not on the protein?
Usually because it was selected and ranked on the peptide. Binding can depend on features that exist only in the synthetic fragment — most commonly the free C-terminal carboxyl. In a published panel of methylation-specific clones, amidating the peptide C-terminus nearly abolished binding for the two clones that had failed on Western blot, and left the two working clones unchanged.
Can hybridoma produce a good PTM-specific antibody?
Yes, but the odds are set before screening starts. Hybridoma and B-cell cloning sample on the order of 104–105 cells and cannot apply selective pressure after the fact, so recovering a rare discriminating clone depends on it happening to be in the panel.
Does a higher affinity clone perform better?
Not reliably. In one published panel the 0.90 nM clone gave only a faint band on full-length protein while the 54 nM clone worked cleanly. In another, the clone binding ~300-fold more tightly to the phosphopeptide was the less selective of the two. Rank by performance in the assay that matters and treat affinity as one input.
How does subtractive selection actually work?
An excess of the soluble unmodified competitor is present during the binding step of panning, so phage that cannot discriminate between the modified and unmodified forms are out-competed and washed away. The discrimination is enforced by the selection, before any clone is picked.
How much starting material does a phage campaign need?
Less than a screening-based route, because the enrichment happens in the library rather than through the number of clones handled. A library built from one immunised animal’s repertoire already contains the diversity; panning extracts it.
Have a site that other platforms could not resolve?
Tell us the modification, the sequence context, and the assay the antibody has to survive in. A PhD scientist reads it and replies within two business days. No NDA needed to start.
Related: phospho / PTM-specific antibody generation · rabbit monoclonal discovery (RabWiz™) · membrane and difficult-antigen discovery · all peer-reviewed publications