Protein pI and pKa Scales: Bjellqvist, Lehninger, and EMBOSS

The quick answer

Bjellqvist, Lehninger, and EMBOSS are not different units. They are alternative sets of pKa assumptions used with the same charge equation. Changing the set can change calculated net charge, the charge-versus-pH curve, and theoretical pI.

The selected pKa set does not change protein length, average molecular weight, or monoisotopic molecular weight. Those results come from the amino acid mass calculation, not the ionization model.

What pKa means in a protein charge calculation

An ionizable group's pKa is the pH where the model treats that group as half protonated and half deprotonated. The calculator uses the Henderson–Hasselbalch relationship to estimate the fractional charge of each ionizable group at a selected pH.

The calculation includes:

• the positively charged N-terminus
• the negatively charged C-terminus
• acidic side chains: aspartate (D), glutamate (E), cysteine (C), and tyrosine (Y)
• basic side chains: histidine (H), lysine (K), and arginine (R)

At low pH the sequence is more protonated and usually more positive. As pH rises, ionizable groups lose protons and the calculated charge becomes more negative. The theoretical pI is the pH where the calculated net charge crosses zero.

pKa values used by Calcorium

The table lists the fixed or default pKa values used by the calculator. Bjellqvist applies additional sequence-specific terminal values described immediately below the table.

How the full Bjellqvist terminal rules work

Bjellqvist is different from the other two options because selected first and last amino acids change the pKa assigned to the terminal group. Calcorium uses the complete terminal-residue rules rather than one fixed value for every sequence.

For other terminal residues, Bjellqvist uses the defaults of 7.50 for the N-terminus and 3.55 for the C-terminus. These terminal adjustments are also used by established Bjellqvist implementations such as Biopython's isoelectric-point module.

Why short proteins and peptides can differ more

Every linear sequence has one N-terminus and one C-terminus, whether it is seven residues or seven hundred residues long. In a short sequence, those two groups can represent a large fraction of all ionizable groups. A different terminal pKa can therefore move the predicted charge and pI noticeably.

Short sequences also contain fewer ionizable side chains. If a peptide has only one histidine, cysteine, lysine, or acidic residue, the chosen pKa for that individual group may have a large influence. In a larger protein, many charged side chains often make the relative contribution from either terminus smaller and produce a smoother charge curve.

This is a tendency, not a rule. Large proteins with unusual compositions, many residues that ionize near the predicted pI, or strongly acidic or basic sequences can still differ meaningfully between pKa sets. ExPASy also notes that poor buffer capacity can make theoretical pI predictions for small proteins problematic.

Example: same composition, different terminal order

The following short artificial sequences contain exactly the same residues. Only their order changes:

• PAAAAAD begins with proline and ends with aspartate.
• DAAAAAP begins with aspartate and ends with proline.

Lehninger and EMBOSS give identical results for this pair because their fixed pKa values depend on residue counts, not terminal residue identity. Bjellqvist distinguishes the sequences because proline changes the N-terminal value and C-terminal aspartate changes the C-terminal value.

What each option is useful for

Bjellqvist

Bjellqvist was developed from polypeptide focusing behavior in immobilized pH gradients and is used by ExPASy Compute pI/Mw. It is a practical default for sequence-based protein pI estimates, proteomics, and comparisons with ExPASy-style results. It is the default in the Calcorium calculator.

Lehninger

The Lehninger option provides a fixed, textbook-style biochemical pKa set. It is useful for teaching, hand-calculation comparisons, and checking how a general reference set changes the estimate. It does not apply residue-specific terminal corrections.

EMBOSS

EMBOSS uses the traditional values from its Epk.dat data file. Select this option when you want a result comparable with an EMBOSSiep or related workflow. The EMBOSS method also assumes that electrostatic interactions do not shift the ionization behavior of individual groups.

Which pKa set should you choose?

• Use Bjellqvist as a practical default or when comparing with ExPASy-style predictions.
• Use EMBOSS when reproducing an EMBOSS analysis or matching an established EMBOSS pipeline.
• Use Lehninger for a general textbook-style estimate or educational comparison.
• Compare all three when a short sequence, unusual composition, or result near a decision boundary makes sensitivity important.

When recording a theoretical pI or net-charge estimate, report the sequence analyzed, the selected pKa set, and whether the sequence represents the mature protein or an unprocessed precursor.

Why theoretical and experimental pI can disagree

All three options are sequence-based estimates. A real residue's pKa can shift because of its local three-dimensional environment, neighboring charges, solvent exposure, ionic strength, temperature, and solvent composition.

The calculator does not automatically model:

• protein folding and residue accessibility
• disulfide-bonded cysteines
• phosphorylation, acetylation, glycosylation, or other modifications
• modified or blocked N- and C-termini
• signal peptide or initiator-methionine removal
• tags, cleavage products, complexes, and multimeric states unless represented in the entered sequence

The result should therefore be reported as a theoretical pI, not treated as an exact experimental value.

Frequently asked questions

Does changing the pKa set change molecular weight?

No. The pKa set changes theoretical charge and pI only. Average and monoisotopic molecular weights are calculated separately from residue masses.

Is one pKa set always the most accurate?

No. Accuracy depends on the sequence, experimental conditions, modifications, and what reference workflow you need to reproduce. Comparing multiple sets is useful for seeing whether the prediction is sensitive to the assumptions.

Why can two sequences with the same composition have different Bjellqvist pI values?

Bjellqvist considers the identity of selected N- and C-terminal residues. Rearranging the same amino acids can therefore change the terminal pKa values and the resulting pI.

References and implementations

• ExPASy Compute pI/Mw documentation
• Biopython Bjellqvist implementation and references
• EMBOSS iep documentation
• IPC: comparison of sequence-based pI methods and pKa sets