Circular Dichroism and Peptide Secondary Structure

Every analytical method discussed on a peptide certificate answers a question about composition: what the molecule weighs, what order the residues are in, how much of the sample is the target. Circular dichroism answers a different question entirely — what shape the chain is holding in the solution it is sitting in.

The principle in one paragraph

Circularly polarised light comes in left-handed and right-handed forms. A molecule built from chiral units absorbs the two unequally, and the difference in absorbance, plotted against wavelength, is the spectrum. A peptide backbone is a chain of chiral centres arranged in a repeating geometry, so the sign and size of that difference depend on the geometry — which is the definition of secondary structure.

A racemic mixture gives no signal at all, because the two enantiomers cancel. This is worth holding onto: the technique exists because peptides are chiral, and it is directly sensitive to the kind of stereochemical change described in racemisation and chiral purity.

Two regions, two different measurements

• Far ultraviolet, roughly 190 to 250 nanometres. The absorbing group is the backbone amide. This region reports secondary structure and is what almost every peptide measurement uses.
• Near ultraviolet, roughly 250 to 320 nanometres. The absorbing groups are the aromatic side chains and any disulfide bonds. This region reports the environment those groups sit in, which makes it a tertiary structure probe — and therefore near-useless for a short peptide, which has no tertiary structure to report on.

Near-ultraviolet spectra also need considerably more material, because the aromatic signals are an order of magnitude weaker than the backbone ones.

What the standard shapes look like

Three reference shapes cover most of what is seen:

• Helix. Two negative bands of similar depth near 208 and 222 nanometres, with a strong positive band near 192. The 222 band is the one usually tracked, because it is least contaminated by other contributions.
• Sheet. A single broad negative band near 216 to 218 nanometres and a positive band near 195. Weaker and more variable than the helix signature, because sheet geometry varies more.
• Disordered chain. A deep negative band near 198 nanometres and little else, sometimes with a shallow negative shoulder around 220.

Real spectra are combinations. Deconvolution algorithms fit an observed spectrum as a weighted sum of reference spectra and return percentages, and those percentages are a fit result rather than a measurement — two structurally different mixtures can fit the same curve.

The measurement is an average over the population

This is the limitation that matters most and is most often forgotten. The spectrum is the sum of every molecule in the cuvette. A sample reported as thirty percent helical may be thirty percent of chains fully helical and seventy percent disordered, or every chain helical along thirty percent of its length, or anything between. The spectrum cannot separate those.

It also cannot say which residues are structured. There is no positional information in a circular dichroism spectrum at all — that belongs to nuclear magnetic resonance or crystallography, and for most synthetic peptides to neither, because neither is run.

What the sample has to satisfy

Far-ultraviolet work is unforgiving about what else is in the cuvette, because anything that absorbs in that region raises the noise until the signal disappears:

• Chloride absorbs strongly below 200 nanometres. Phosphate or fluoride buffers at low concentration are the usual substitutes.
• Trifluoroacetate absorbs in the same region. A peptide supplied as a trifluoroacetate salt carries a counter-ion that interferes with its own structural measurement, which is one of the practical reasons for salt exchange — the trade-off is set out in counter-ions and salt form.
• Concentration must be known accurately, because the reported quantity is normalised per residue. An error in concentration propagates directly into the reported helix content, so the determination described in concentration by A280 is part of the measurement, not preparation for it.
• Aggregates scatter light and distort the baseline. A spectrum from a hazy solution is not a structural result, and the behaviour that produces the haze is covered in peptide aggregation in solution.

Temperature, and what a melt curve is

Tracking one wavelength — usually 222 nanometres — while the temperature rises produces a curve that falls as the structure unfolds. The midpoint of that transition is reported as an apparent melting temperature.

Apparent is the operative word. It is a property of the peptide in that buffer at that concentration, not a constant of the molecule, and it is only meaningful as a thermodynamic value if the transition reverses on cooling. Many peptides aggregate on heating instead of unfolding cleanly, giving an irreversible curve whose midpoint is an aggregation onset rather than an unfolding midpoint.

Why most short peptides give a disordered spectrum

A chain needs roughly three to four turns of helix before the pattern is stable, and a short peptide in plain aqueous buffer usually has too little of the chain to hold any geometry against thermal motion. The honest result for many synthetic peptides is the disordered signature, and that is not a defect in the material.

Structure is frequently induced rather than intrinsic — adding trifluoroethanol or a membrane-mimetic detergent will produce a helical spectrum from a chain that is disordered in buffer. Such a result is a statement about the peptide’s propensity under those conditions, and reporting it without the conditions is meaningless. Fragments with this behaviour are common; one example is discussed in the analysis of the LL-37 cathelicidin fragment.

Where it belongs among the other methods

Circular dichroism is not an identity test and not a purity test. It will not distinguish a peptide from a deletion sequence, it will not detect a residual solvent, and a highly impure sample can give a clean spectrum if the impurities are unstructured.

What it does is answer a question no chromatogram or mass spectrum reaches — an independent one, in the sense described in confirming identity with two independent methods: whether a correctly synthesised chain has folded the way it is supposed to. For most short peptides that question does not arise, which is why the method is absent from routine certificates. It becomes relevant with disulfide-constrained and longer chains, where correct mass and correct connectivity still leave folding open — the situation set out in where peptide analysis becomes protein analysis.