How we read an HPLC chromatogram of a peptide

A chromatogram is a simple drawing. A nearly flat line, one tall peak, a few bumps. People who meet one for the first time in a certificate usually skip straight to the percentage printed underneath, which we think is a pity, because the trace says more than the figure does and, at the same time, less than it seems to. Here we go through what can be seen on it when the sample is a research peptide.

What the two axes show

The horizontal axis is time, in minutes, counted from the moment the sample was injected onto the column. The vertical axis is the detector response, meaning ultraviolet absorbance, usually given in milli-absorbance units (mAU).

For peptides the usual method is reversed-phase HPLC (RP-HPLC). The column is packed with silica carrying bonded C18 chains, which are hydrophobic. The mobile phase is a mixture of water and acetonitrile with a little trifluoroacetic acid (TFA), and the proportion of acetonitrile rises steadily over the run. That is what we call the gradient. More hydrophobic molecules hold on to the column more tightly and come off later, once the organic solvent is concentrated enough to release them.

The detector reads at 214–220 nm. That is where the peptide bond itself absorbs, so any chain of amino acids gives a signal. At 280 nm only sequences containing tryptophan or tyrosine would respond, and the rest would pass unseen.

The main peak and its retention time

The tallest peak is normally the target peptide. The point at which it reaches its maximum is called the retention time.

Retention time is not a property of the peptide. It is a property of the peptide in that method: the same column, the same gradient, the same temperature, the same flow rate. Change the slope of the gradient and the peak moves. So there is no point comparing the minute on one supplier's chromatogram with the minute on another's. What does make sense is setting two batches side by side when both were analysed with the same method. If the main peak turns up in the same place, that is a first sign of consistency.

We look at the shape too. A narrow, symmetrical peak points to a clean separation. A “shoulder” on one side can mean that two substances sit under the peak, eluting almost together. A long tail is harder to judge, because sometimes it comes from the column and not from the sample.

The small peaks next to it

Solid-phase synthesis adds amino acids one at a time, and no step goes to completion for every molecule. What is left behind are chains missing one amino acid, chains that stopped before the end, chains still carrying a protecting group that never came off. On top of those come the degradation products, such as oxidation of methionine or deamidation of asparagine.

All of them resemble the target peptide chemically, so they elute close to it, a little before or a little after. A convenient example is a short peptide with methionine in its sequence, like the one on the Semax product page in the PEPTIDÁE catalogue (their site is in Romanian). For any molecule of this kind, the oxidised form is more polar and generally appears slightly earlier than the main peak.

Two features of the trace are not impurities. The first is the signal in the opening minutes, where the sample solvent and everything the column does not retain come through. The second is the baseline drifting gently up or down across the gradient, because TFA also absorbs at short wavelengths and the composition of the mobile phase changes from one minute to the next. As a rule, neither enters the calculation.

How the percentage comes out of the areas

The chromatography software draws a baseline under each peak and calculates the area between the peak and that line. HPLC purity is the area of the main peak divided by the sum of the areas of all integrated peaks, multiplied by a hundred. Below is an example we made up ourselves, with round numbers, purely to show the arithmetic.

The reported purity would be 98.2%. The method is called area percent, and it rests on one assumption: that every component absorbs equally well at the chosen wavelength. For impurities related to the peptide the assumption is reasonable, since they have almost the same number of peptide bonds. Exact it is not.

The figure also depends on how the integration was done. A higher detection threshold leaves the very small peaks out, and purity goes up on paper without the sample having changed at all. A gradient that is too steep crowds the impurities under the main peak and hides them. When we have two percentages to compare, we first check whether the methods are alike.

What the chromatogram does not say

It does not say which substance is in the peak. Retention time shows how hydrophobic the molecule is under the given conditions, nothing more. Two different peptides can come off in the same minute. Identity is confirmed by mass spectrometry, which measures the molecular mass and compares it with the one calculated from the sequence. How the two methods complement each other is the subject of a post on the same site's blog, about HPLC and mass spectrometry in purity control.

Nor does it say how much peptide is in the vial. The 98.2% in our example means that, of everything that absorbed in the ultraviolet and was integrated, 98.2% is the main peak. The water left after lyophilisation and the counter-ions that accompany the peptide as a salt do not appear among the counted peaks. They have mass all the same. Net peptide content is determined separately, by other analyses, and is generally lower than the purity.

And it stays silent about anything that gives no signal at the detector, such as inorganic salts.

When the chromatogram arrives without context

A cropped image with no header is of little help to us. A complete chromatogram carries the sample name or batch number, the run date, the wavelength and, below the trace, the peak table with retention times and areas. Without the table, the reader cannot rebuild the percentage. Without the batch number, we do not know whose sample it is.

We have a simple habit: we print the chromatogram and write the column, the gradient and the wavelength in the margin. With two sheets side by side, from different batches run with the same method, the comparison takes a minute.