The Gold Standard of Biochemistry | Protein A280, Extinction Coefficients and Turbidity Bias

A protein can absorb ultraviolet radiation around 280 nanometres because its aromatic amino acids and certain disulfide bonds interact with light. When the sample is a purified and identified protein, an A280 measurement can help quantify the protein quickly. Yet the instrument directly supplies an optical response rather than a count of protein molecules or a protein mass concentration.

The same mass concentration of two different proteins can give different readings because their amino-acid sequences differ. A protein containing several tryptophan residues may be highly responsive at 280 nm, while one with few aromatic residues produces a relatively weak signal. This is why a universal conversion of one optical absorbance unit to one milligram per millilitre is scientifically unreliable.

The Harvard Center for Macromolecular Interactions describes how to calculate a sequence-based extinction coefficient and advises a suitable absorbance range, replicate readings and matched buffer blanks. The ExPASy ProtParam documentation gives a standard estimate based on tryptophan, tyrosine and cystine counts. Protein identity therefore enters the calculation before an absorbance can become a molar concentration.

A sample may also contain nucleic acid, additives, particles or aggregates. Some can absorb UV radiation; others can scatter it. The instrument can register a signal that is partly unrelated to absorption by the intended protein. A large, precise-looking A280 number is not proof of a pure, intact or functional protein.

The correct chain is: identify the sample and optical environment, obtain or justify ε280, verify the path length, correct against a suitable blank, then calculate concentration and report the assumptions. Each step may contribute uncertainty or cause systematic bias.

This article uses an imaginary 50-kDa protein to make that chain checkable. It deliberately separates instrument physics, sequence-based chemistry and protein function.

Contents · Next chapter

ExPASy ProtParam estimates the molar extinction coefficient of a protein measured in water near 280 nm using the expression ε280=5500NTrp+1490NTyr+125Ncystine. The coefficient is in M⁻¹ cm⁻¹, and the N quantities count tryptophan residues, tyrosine residues and disulfide-linked cystine pairs respectively.

Imagine a purified teaching protein with three tryptophan residues, six tyrosine residues and two cystine pairs. Its tryptophan contribution is 3×5500=16,500; its tyrosine contribution is 6×1490=8,940; and its cystine contribution is 2×125=250 M⁻¹ cm⁻¹.

Adding them gives ε280=25,690 M⁻¹ cm⁻¹. This number is an estimate for that fictional sequence under the model, not a measured molar absorptivity for an actual sample. It should be revisited if the protein contains a covalently linked chromophore, chemically modified aromatic groups or other unusual optical features.

Cystine is not simply another word for individual free cysteine residues. A cystine contribution is associated with a disulfide bond joining two cysteine-derived groups. ProtParam provides assumptions for oxidised and reduced cysteine states, and a real protein’s disulfide state may need verification.

The documentation also cautions that extinction-coefficient estimates can be more uncertain for proteins lacking tryptophan. Thus even this chemically informed calculation should not be presented as exact under every solvent and conformation.

A meaningful report includes protein identity, sequence or measurement source of the coefficient, disulfide assumption, wavelength, solvent and whether a direct experimental determination was made.

Contents · Previous chapter · Next chapter

For a homogeneous sample under valid Beer–Lambert conditions, A=εcl. Absorbance A is dimensionless, ε is in M⁻¹ cm⁻¹, concentration c in mol/L, and path length l in centimetres. The dimensional terms cancel, which makes unit checking an excellent first quality control.

Give the imaginary protein a molar mass of 50,000 grams per mole. A concentration of 1.00 mg/mL equals 1.00 gram per litre. Therefore c=(1.00 g/L)/(50,000 g/mol)=2.00×10⁻⁵ mol/L=20.0 µM.

With ε=25,690 M⁻¹ cm⁻¹ and a one-centimetre path, A=(25,690)(2.00×10⁻⁵)(1)=0.5138. If the sample were diluted twofold with no chemical change, the expected concentration would be 0.50 mg/mL or 10.0 µM, and the ideal A280 would be 0.2569.

Conversely, a properly blanked A280 of 0.5138 for this exact protein and path yields c=A/(εl)=2.00×10⁻⁵ M. Multiplying by its molecular mass recovers 1.00 g/L or 1.00 mg/mL.

Do not confuse a mass concentration with a molar concentration. A different protein at 1.00 mg/mL may have a different molarity because its molecular mass differs. A protein forming an oligomer also requires clarity about whether molar concentration refers to monomer units or complete assemblies.

The law describes an assumed linear optical regime. Instrument detector range, stray radiation, scattering and changes in protein chemistry can make high or low measured absorbances unsuitable for this simple conversion.

Contents · Previous chapter · Next chapter

Aggregates, particles or droplets in a protein sample may scatter light. The amount of radiation reaching the detector can change even if the absorbing protein concentration remains constant. Depending on optical geometry, this can be interpreted as additional optical density and can produce a concentration estimate larger than the actual protein amount.

Use a transparent numerical sensitivity exercise rather than claiming one universal scattering law. Suppose the fictional 1.00 mg/mL sample should give an intrinsic A280=0.5138 but has an additional apparent absorbance-equivalent signal of +0.1000. The displayed value becomes 0.6138.

A naive Beer–Lambert conversion gives 0.6138/(25,690×1)=2.389×10⁻⁵ M, equivalent to 1.1946 mg/mL at the assumed 50-kDa molecular mass. The reported concentration is roughly 19.46% above the actual fictional concentration.

Now imagine a 0.50 mg/mL preparation with intrinsic A280=0.2569 but the same added apparent background. It would read 0.3569 and falsely imply approximately 0.6946 mg/mL. Its relative error is almost 39%, showing why a fixed apparent optical offset has a greater relative effect at low signals.

Real particle scattering does not necessarily contribute a constant additive absorbance at every concentration. Particle number, size, aggregation state, wavelength, blank composition and detector optics all matter. Our +0.1000 term is specifically an invented absorbance-equivalent error, not a validated correction formula.

If the measured result is surprising, inspect the sample and an appropriate UV spectrum, check blanking, and compare measurements across suitable dilutions or an independent protein quantification method. A clear-looking tube can still contain small particles; visual inspection alone is not sufficient proof of a pure monomeric protein.

Contents · Previous chapter · Next chapter

The same protein has a smaller measured absorbance when the optical path is shorter under Beer–Lambert assumptions. Our 1.00 mg/mL protein is predicted to give A280=0.5138 in one centimetre but A280=0.05138 in a 0.10-centimetre path, because A is proportional to l.

A microvolume spectrophotometer may vary the light path or report a value normalised to one centimetre. Its displayed software output and its raw detector data are not necessarily expressed on the same path-length basis. Applying a second path-length correction to an already normalised value can misstate concentration badly.

A good blank contains the relevant buffer and solvent components without the analyte. Certain reducing agents and other additives absorb in the UV; Harvard’s guidance specifically cautions about DTT-containing buffers. A water blank cannot automatically remove the UV signal of a non-water sample matrix.

Dilution testing can expose failures of linearity. If a sample is measured at 1.00, 0.50 and 0.25 mg/mL under otherwise consistent ideal conditions, A280 should scale as 0.5138, 0.2569 and 0.12845. Back-calculated concentrations should agree after the recorded dilution factors are restored.

However, sample dilution can alter aggregation, adsorption, equilibria or protein stability. A dilution series should therefore be treated as an experiment with its own assumptions, not as an automatic cure for every instrument or chemical problem.

Instrument performance checks should confirm appropriate wavelength accuracy, detector range, path-length response and reference stability. Replicate readings, drift controls and a credible uncertainty estimate matter more than the number of decimals displayed on the instrument.

The broader analytical lesson is that sample optics and chemistry must correspond to the same model used to infer concentration.

Contents · Previous chapter · Next chapter

Suppose the fictional protein loses one tryptophan in a hypothetical variant while all other counted chromophores remain unchanged. Its sequence-based coefficient would decrease by 5500 from 25,690 to 20,190 M⁻¹ cm⁻¹.

Keep an assumed true molar concentration of 20.0 µM and a one-centimetre optical path. The expected absorbance for this variant becomes A280=20,190×20.0×10⁻⁶=0.4038, rather than the original protein’s 0.5138.

If someone processes the variant’s A280=0.4038 using the old 25,690 coefficient, they infer about 15.72 µM instead of 20.0 µM. Under a deliberately fixed molecular-mass assumption for this sensitivity exercise, that is an underestimation of concentration by approximately 21.4%.

In actual molecular biology, replacing a tryptophan changes molecular mass and may change folding or environment. A real measurement must update all applicable molecular properties rather than freeze them as in this simplified comparison.

The example also makes clear why protein purity and enzymatic activity cannot be inferred from the UV coefficient. A denatured protein can still contain its aromatic residues and absorb near 280 nm even when its catalytic function has been destroyed.

For an unknown mixture of proteins, a single sequence coefficient cannot be assumed without a valid mixture model. A280 is most interpretable when the chemical identity and composition of the preparation are established.

Contents · Previous chapter · Next chapter

Does A280=1.0 mean 1.0 mg/mL protein? No. The appropriate extinction coefficient, molecular mass and path length are needed.

Why do tryptophan and tyrosine contribute? Their aromatic groups absorb UV radiation in this region, influencing the protein-specific molar absorptivity.

Does each cysteine add the cystine coefficient? No. ProtParam’s cystine term is associated with disulfide-linked cysteine pairs.

Can a protein without tryptophan be measured by A280? Sometimes, but its response can be weak and estimates may be less reliable.

Does an elevated A280 prove more protein was made? No. Other absorbers, scattering, path length and instrument changes can produce different readings.

Can a matched blank matter even when absorbance is high? Yes. A sample buffer can absorb near 280 nm; a water-only blank may leave an unwanted contribution.

Can the same protein give a different A280 in a microvolume instrument? Yes, if the optical path or reported path-length convention differs.

Does A280 establish that the enzyme is active? No. Functional enzyme fraction and total protein amount are different measurements.

Does a good 260/280 ratio establish perfect protein purity? No. A single spectral ratio cannot exclude every contaminant or mixture.

Should scattering always be subtracted as a fixed 0.1 absorbance? No. The number in this article is a fictional sensitivity input, not a correction rule.

Contents · Previous chapter · Next chapter

Accurate protein concentration can become the denominator for enzyme-specific activity or turnover calculations. If the protein amount is overestimated, an activity expressed per milligram may appear too low even when the product-formation rate was measured correctly. Conversely, correct total protein concentration does not prove every molecule is functioning.

Reconstruct the original calculation: the fictional ε is 25,690 M⁻¹ cm⁻¹, molecular mass is 50,000 g/mol and a concentration of 1.00 mg/mL corresponds to 20 µM and A280≈0.5138 at one centimetre. Repeat after changing path length and sequence count. Explain which part of the model changes and which stays fixed.

Continue with The Gold Standard of Biochemistry, Active Enzyme Fraction and Turnover, Coupled Assays and the Reported Rate, Beer–Lambert Calibration and Stray Light, Learning Transfers into Life, and How X Works.

Authoritative sources reviewed on 26 October 2026: ExPASy ProtParam Documentation, Harvard Center for Macromolecular Interactions: Measuring Concentration, and Evaluation of Peptide and Protein Aggregation by Spectroscopic Methods.

All numerical protein and interference examples in this guide are synthetic. The additive absorbance-equivalent background is a teaching assumption, not a validated method for correcting turbidity or instrument optical density.

Contents · Previous chapter · Biochemistry hub