Short answer: The molar extinction coefficient (ε), also called the molar absorptivity or molar absorption coefficient, is the constant in the Beer-Lambert law, A = εcl, that links a solution's absorbance (A) to its molar concentration (c) and the light path length (l) [1][5]. It is usually quoted in M⁻¹ cm⁻¹ [1]. For peptides and proteins measured at 280 nm, ε can be estimated from the sequence alone, because absorbance at that wavelength comes from tryptophan, tyrosine and cystine (disulfide-bonded cysteine pairs): ε280 ≈ (number of Trp × 5,500) + (number of Tyr × 1,490) + (number of cystines × 125) M⁻¹ cm⁻¹ [2][4]. Sequences without Trp or Tyr have no usable 280 nm signal, so laboratories turn to the far-UV, where the peptide bond itself absorbs [7][8].
This article is educational analytical chemistry for laboratory readers. It does not describe effects in people or animals and is not guidance for human or animal use.
The Beer-Lambert law in one paragraph
A spectrophotometer compares the intensity of light passing through a reference (I₀) with the intensity passing through the sample (I) and reports absorbance as A = log₁₀(I₀/I) [1]. Absorbance is proportional both to concentration and to the path length, normally the width of the cuvette; the proportionality constant is ε, giving A = εcl [1]. Rearranged, ε = A/(lc): the absorbance a 1 mol/L solution would give over a 1 cm path [1]. Because A is logarithmic, an absorbance of 1 means 90% of the light at that wavelength was absorbed and 10% transmitted [1].
For a purified protein or peptide, absorbance spectroscopy is described by Grimsley and Pace as the most convenient and accurate way to measure concentration in solution [5]. Once ε is known, concentration is simply c = A/(εl).
Names and units you will see
- Molar extinction coefficient, molar absorptivity, molar absorption coefficient. These are used interchangeably in the literature for the same ε [1][2][7]. Pace and co-workers and Grimsley and Pace use "molar absorption coefficient" [2][5].
- Molar units: M⁻¹ cm⁻¹. With concentration in mol/L and path length in cm, the unit is mol⁻¹ dm³ cm⁻¹ (equivalently M⁻¹ cm⁻¹), and tables often omit it [1].
- Mass units: mL mg⁻¹ cm⁻¹. Dividing a molar ε by the molecular weight gives the absorbance expected for a 1 mg/mL solution. ExPASy's ProtParam documentation states this as Absorb(Prot) = E(Prot) / Molecular_weight [4], and Anthis and Clore convert between the two forms the same way [7].
Why 280 nm: tryptophan, tyrosine and cystine
Absorbance at 280 nm is highly specific because it arises strictly from tryptophan and tyrosine side chains, plus a small contribution from disulfide bonds if present [7]. ProtParam's documentation adds that free cysteine does not absorb appreciably at wavelengths above 260 nm, whereas cystine does [4]. (For background on how cystine forms, see our guide to the disulfide bond; residue abbreviations are in our amino acid codes reference.)
That specificity is what makes sequence-based prediction possible: if you know how many of each chromophore a chain contains, you can add up their contributions [3][7].
Where the residue values come from
The coefficients in common use were refined over several decades:
- Edelhoch (1967) model compounds. Edelhoch measured Trp and Tyr values using blocked amino acid analogues — N-acetyl-L-tryptophanamide for Trp and glycyl-L-tyrosylglycine for Tyr — in 6.0 M guanidine hydrochloride, 0.02 M phosphate buffer, pH 6.5 [2][4].
- Gill and von Hippel (1989). They presented a method for calculating molar extinction coefficients at 280 nm from amino acid composition, calibrated against 18 "normal" globular proteins with accurately known coefficients, and reported accuracy of ±5% in most cases [3]. ProtParam's documentation notes that 6 of the 44 values in that comparison deviated by more than 10% [4].
- Mach, Middaugh and Lewis (1992). Using matrix linear regression against experimentally determined coefficients for more than 30 proteins, they derived 5,540 M⁻¹ cm⁻¹ for tryptophan and 1,480 M⁻¹ cm⁻¹ for tyrosine in an "average" protein, and reported a significant improvement over the model-compound values measured in guanidine-HCl [6].
- Pace et al. (1995). From 116 measured coefficients for 80 proteins, they concluded that ε280 for a folded protein in water is best predicted as (#Trp)(5,500) + (#Tyr)(1,490) + (#cystine)(125) M⁻¹ cm⁻¹ [2]. They found that the average Trp value in proteins is lower than the value measured in any of the solvents they tested, while the average Tyr value falls between those measured in 6 M guanidine hydrochloride and in propanol [2].
How to estimate ε280 from a peptide sequence
1. Count the chromophores. Tally Trp (W) and Tyr (Y) residues, and the number of disulfide bonds (cystines) [2][4].
2. Apply the coefficients. ε280 = (#Trp × 5,500) + (#Tyr × 1,490) + (#cystine × 125) M⁻¹ cm⁻¹ [2][4].
3. Decide how to treat cysteines. ProtParam reports two numbers: one assuming every pair of Cys residues forms a cystine, and one assuming all Cys residues are reduced [4]. Choose the value that matches the oxidation state of the sample.
4. Convert to mass units if needed. Divide ε by the molecular weight to get the absorbance of a 1 mg/mL solution [4].
Worked example (arithmetic only). A hypothetical peptide containing one Trp, two Tyr and no disulfide has ε280 = 5,500 + (2 × 1,490) = 8,480 M⁻¹ cm⁻¹ using the coefficients above [2]. If a solution of it reads A280 = 0.424 in a 1 cm cuvette, c = 0.424 / 8,480 = 5.0 × 10⁻⁵ M, or 50 µM. If the peptide's molecular weight were 2,000 g/mol, the mass-unit coefficient would be 8,480 / 2,000 = 4.24 mL mg⁻¹ cm⁻¹, and 50 µM would correspond to 0.1 mg/mL.
A real example from the literature: Anthis and Clore list a 22-residue peptide from the calmodulin-binding domain of CaMK1 and the 26-residue skMLCK M13 peptide, each with a predicted ε280 of 5,500 M⁻¹ cm⁻¹ from a single tryptophan [7].
How reliable is the 280 nm estimate?
- Trp-containing sequences predict best. Pace et al. describe their ε280 values as quite reliable for proteins containing Trp and less reliable for those that do not [2]. ProtParam warns that error can exceed 10% for proteins without Trp [4].
- Measuring beats predicting. Pace et al. concluded that the Edelhoch method is the best way to measure ε for a protein and that the best approach is to measure rather than predict it [2].
- The context matters. Gill and von Hippel discuss the assumptions behind their method and its limitations, and Mach et al. note that experimentally determined coefficients themselves can vary considerably [3][6].
- The solution must be clean at 280 nm. Most common buffers are essentially transparent at 280 nm, whereas many solutes absorb in the far-UV [7].
Why not simply weigh the peptide?
Weighing a dried sample sounds more direct, but Anthis and Clore point out that calibrating by directly weighing lyophilized protein or peptide is confounded by any residual salt or water left in the dried material [7]. Assays that do not rely on intrinsic chromophores, such as Bradford or Lowry, also need calibration plus extra reagents and time [7]. This is the same issue behind the difference between gross weight and peptide content discussed in our article on net peptide content vs purity, and a reason the dried weight of lyophilized peptides is not a direct measure of how much peptide is present [7].
Peptides with no Trp or Tyr: the far-UV options
If a sequence contains no tryptophan or tyrosine, A280 cannot be used to determine its concentration [7]. Grimsley and Pace include a sensitive method for proteins containing few if any Trp or Tyr residues [5]. Two published approaches use the backbone peptide bond, which absorbs in the far-UV:
205 nm (Anthis and Clore, 2013).
- Absorbance at 205 nm arises primarily from the peptide backbone, although a protein's absorbance maximum actually lies closer to 190 nm; 205 nm is favoured partly because measuring at lower wavelengths is technically harder [7].
- A widely used generic value is ε205 = 31 mL mg⁻¹ cm⁻¹, which needs no sequence information [7].
- Anthis and Clore instead sum sequence-specific contributions, using a backbone value of 2,780 ± 168 M⁻¹ cm⁻¹ per peptide bond plus side-chain values that include Trp 20,400, Phe 8,600, Tyr 6,080, His 5,200, Met 1,830, Arg 1,350, Cys 690 and Asn and Gln 400 M⁻¹ cm⁻¹; for each disulfide they add 820 M⁻¹ cm⁻¹ [7].
- Across eight test polypeptides, their sequence-specific values gave concentration errors of at most 8.2% compared with A280-derived concentrations, versus up to 19.0% with the generic 31 mL mg⁻¹ cm⁻¹ value and up to 14.2% with an earlier ratio-based (Scopes) formula [7].
- They report that the method works for both folded proteins and short unstructured peptides, and that side chains can account for up to 40% of total A205, which is why sequence-specific values help [7].
- On average A205 is about 30 times A280, giving higher sensitivity, but many buffer components absorb at 205 nm, so buffer-only blanks must be measured and subtracted [7].
- As a demonstration, they quantified the N-terminal domain of calmodulin (residues 1–76), which has no residues that absorb at 280 nm, using a calculated ε205 of 266,150 M⁻¹ cm⁻¹ [7].
- Kuipers and Gruppen measured coefficients for the 20 amino acids and the peptide bond at 214 nm, in the presence of acetonitrile and formic acid, to allow quantitative comparison of peptides separated by reversed-phase HPLC and identified by mass spectrometry [8].
- They report 923 M⁻¹ cm⁻¹ for the peptide bond; tryptophan absorbs about 30 times more than a peptide bond, phenylalanine, tyrosine and histidine about six times more, and methionine about the same [8].
- Proline is negligible as a free amino acid but, when inside a peptide chain (other than at the N-terminus), absorbs about three times more than a peptide bond [8].
- Calculated and measured coefficients for several validation proteins and peptides were mostly in good agreement [8].
Practical checklist
- Confirm the sequence and disulfide state before calculating ε, and pick the oxidised or reduced value accordingly [4].
- Blank against the same buffer, especially in the far-UV, where many solutes absorb [7].
- Keep track of units. Mixing M⁻¹ cm⁻¹ with mL mg⁻¹ cm⁻¹ is a common source of error; convert using molecular weight [4][7].
- Treat predictions as estimates. Expect the best agreement for Trp-containing sequences, and measure ε directly where accuracy matters [2][4].
FAQ
Is the extinction coefficient the same as molar absorptivity?
Yes. ε in A = εcl is called the molar absorptivity or molar extinction coefficient; Pace and co-workers call it the molar absorption coefficient [1][2].
Why does phenylalanine not count at 280 nm?
The 280 nm prediction uses only Trp, Tyr and cystine, because absorbance at that wavelength arises from those chromophores [2][7]. Phenylalanine does absorb significantly in the far-UV, at 205 and 214 nm [7][8].
Does cysteine absorb at 280 nm?
Free cysteine does not absorb appreciably above 260 nm; cystine (a disulfide-linked pair) does, contributing 125 M⁻¹ cm⁻¹ per cystine at 280 nm [2][4].
How do I measure a peptide that has no Trp or Tyr?
Use far-UV absorbance: a sequence-specific ε205 such as the Anthis and Clore method, or 214 nm coefficients such as those of Kuipers and Gruppen [7][8].
What is the difference between molar and mass extinction coefficients?
The molar form (M⁻¹ cm⁻¹) refers to 1 mol/L; the mass form (mL mg⁻¹ cm⁻¹) refers to 1 mg/mL and equals the molar ε divided by molecular weight [4][7].
Bottom line
The molar extinction coefficient turns an absorbance reading into a concentration through A = εcl [1][5]. At 280 nm it can be estimated from a peptide's Trp, Tyr and cystine content using 5,500, 1,490 and 125 M⁻¹ cm⁻¹ respectively, with the best accuracy for Trp-containing sequences [2][4]. For sequences without those residues, far-UV methods based on the peptide bond — sequence-specific ε205 or 214 nm coefficients — fill the gap, provided buffer absorbance is controlled [7][8]. Where accuracy matters, Pace and co-workers' advice stands: measure ε rather than predict it [2].
All products are for research use only — not for human or animal consumption.