Why 280 nm Is the Protein Absorbance Wavelength
Most purified proteins dissolved in water look like water: clear and colourless. Shine ultraviolet light through it, though, and the solution soaks up a band of that light near 280 nanometres. The protein absorbance wavelength of 280 nm comes from a handful of amino acids with ring-shaped side chains. It’s also why one reading on a spectrophotometer, called A280, can tell you how much protein is in the tube without adding any reagent.
Why the protein absorbance wavelength is 280 nm
Only three of the twenty amino acids that make up proteins absorb much in the near ultraviolet: tryptophan, tyrosine and phenylalanine. All three have aromatic rings. An NPTEL university lecture on biological chromophores gives typical figures. Tryptophan absorbs most strongly, with a molar absorption coefficient of about 5,700 M−1cm−1 at 280 nm. Tyrosine reaches about 1,400 at 274 nm. Phenylalanine is weak, about 200, and peaks lower, near 250 to 257 nm according to a table from National Tsing Hua University in Taiwan.
So at 280 nm the signal is almost all tryptophan and tyrosine. There’s one minor extra. When two cysteines are joined by a disulfide bond, the pair (called cystine) absorbs weakly in the 250 to 270 nm region. A single free cysteine, by contrast, doesn’t absorb appreciably above 260 nm. The protein backbone itself absorbs much further into the ultraviolet, around 190 nm for the peptide bond, so it stays out of the way at 280.
That makes 280 nm a practical choice. It sits near tryptophan’s peak, picks up most of tyrosine’s band, and avoids the far-UV region where the backbone and many other side chains absorb.
From A280 to a concentration
Turning a reading into an amount uses the Beer-Lambert law, A = εlc. Absorbance (A) equals the molar extinction coefficient (ε) times the path length in centimetres (l) times the molar concentration (c). Odinity’s lab comparing Bradford, UV and Biuret methods used this relation with bovine serum albumin (BSA) standards and got a straight-line fit at 280 nm with a correlation of 0.9988.
Absorbance is a logarithm of how much light gets through, so it compresses quickly. A biophysics facility guide from the University of Basel’s Biozentrum points out that an absorbance of 1 means 90% of the light is absorbed, and 2 means 99%. Above about 1.5 so little light reaches the detector that readings become unreliable. The same guide warns that drop-based instruments with a 1 mm path usually rescale results to a 1 cm path. For very dilute samples, the instrument is actually reading a tenfold smaller absorbance than it reports, so the result is less accurate than the number suggests.
Getting protein concentration from A280 is simple once you know ε. Rearranged, c = A/(εl). A reading of 0.5 in a 1 cm cuvette, for a protein with ε of 42,925 M−1cm−1, works out to about 12 micromolar. Multiplying by the molecular mass converts that to milligrams per millilitre. The hard part is knowing ε.
Where the extinction coefficient comes from
Because only tryptophan, tyrosine and cystine absorb at 280 nm, a protein’s coefficient can be estimated by counting them. Edelhoch measured the two aromatic residues in 1967 using small model compounds. In 1989 Gill and von Hippel showed those values also worked for native proteins, matching measured coefficients for 18 proteins “to ± 5% in most cases.” Six of their 44 values were off by more than 10%.
In 1995 Nick Pace and colleagues at Texas A&M calibrated the sum against 116 measured coefficients from 80 proteins. Their paper in Protein Science gives the equation most labs now use, for a folded protein in water:
ε280 = (number of Trp × 5,500) + (number of Tyr × 1,490) + (number of cystines × 125), in M−1cm−1.
The ProtParam tool on the Expasy server uses these values. It reports two numbers, one assuming every cysteine pair forms a cystine and one assuming all cysteines are reduced. According to its documentation, the average deviation from measured values was about 3.8% overall and 3.2% for proteins that contain tryptophan. Proteins without tryptophan often missed by more than 10%. Tryptophan contributes much more to the total than tyrosine, and its coefficient is less sensitive to its surroundings.
The BSA molar extinction coefficient as a worked example
BSA is the usual standard protein, so it makes a good test. The UniProt entry for bovine albumin (P02769) lists a mature chain of 583 residues with 2 tryptophans, 20 tyrosines and 35 cysteines forming 17 disulfide bonds. Plugging those into the Pace equation gives 11,000 + 29,800 + 2,125 = 42,925 M−1cm−1. If all the cysteines were reduced, the figure would be 40,800. With a molecular mass of about 66,400, a 1 mg/mL solution should read roughly 0.65.
Pace’s table of 80 proteins lists four published values for BSA at 280 nm, from 41,568 to 43,890; the first of them, 43,890, sits about 2% above the prediction. At 278 nm, Pace’s own measurement by the Edelhoch method gave 44,600, against an average of earlier literature values of 43,820. The Odinity lab report quotes 43,824, which matches that literature figure. These are close but not identical, which is the realistic precision for this kind of number.
Illustration of sample vials, one cloudy, the kind of difference that skews a UV reading (artist’s impression).
The 260/280 problem
The biggest trap in A280 is nucleic acid. The bases in DNA and RNA absorb strongly near 260 nm. The Tsing Hua table lists adenosine monophosphate at 15,000 M−1cm−1 at 259 nm, against 5,000 for tryptophan at 280. Nucleic acid also absorbs at 280 nm, though less strongly than at 260 nm: the Odinity lab’s pure DNA sample read 1.643 times higher at 260 nm than at 280 nm. So leftover DNA or RNA makes a protein sample read high.
The usual check is the ratio of absorbance at 260 nm to absorbance at 280 nm. A 2014 review by Bertrand Raynal and colleagues at the Institut Pasteur says the 260/280 ratio “should be close to 0.57 for a non-contaminated protein sample.” A higher ratio, seen as a bump at 260 nm, points to nucleic acid. In the Odinity lab, pure BSA gave 0.623 and pure DNA gave 1.643.
Older methods try to correct for it. A laboratory protocol chapter gives the formula attributed to Layne (1957): protein in mg/mL = 1.55 × A280 − 0.76 × A260. A 1996 BioTechniques note by Keith Manchester of the University of the Witwatersrand shows that these coefficients assume a protein whose 1 mg/mL solution absorbs about 0.9 at 280 nm, and RNA that absorbs twice as strongly at 260 nm as at 280 nm. For serum albumin, which absorbs about 0.7, the factors would instead be 2.0 and 1.0. Treat such formulas as rules of thumb for a typical protein, not exact corrections.
Raynal’s group adds that common reducing agents such as DTT raise absorbance at 250 nm and below, which distorts the shape of the 280 nm peak. Their advice is to record the whole spectrum from at least 240 to 350 nm, rather than a single wavelength.
Where A280 falls short
Composition. Two proteins at the same mass concentration can give very different readings if one has more tryptophan. Odinity’s BCA protein concentration assay report calls UV at 280 nm inconsistent for exactly this reason. A protein with no tryptophan or tyrosine absorbs very little at 280 nm, since only weak contributions from cystine and, negligibly, phenylalanine remain, so A280 is not a usable measure for it. Such proteins can be measured further into the ultraviolet, for example at 205 nm, where proteins absorb far more strongly because of their peptide bonds, or with a dye-based assay.
Light scattering. Aggregates scatter light, and the instrument counts scattered light as absorbed. The Basel guide gives a correction: subtract twice the absorbance at 333 nm from the reading at 280 nm. The factor of two comes from Pace’s 1995 paper and assumes scattering falls off with the fourth power of wavelength, as in Rayleigh scattering. It notes that a sloping, non-zero baseline above 300 nm is the sign of scattering. Raynal and colleagues likewise say aggregate-free proteins shouldn’t absorb above 320 nm.
Other chromophores. Gill and von Hippel assumed no other light-absorbing groups were present. That rules out proteins with coloured cofactors, such as hemoglobin, catalase and peroxidase.
Purity. The Odinity methods lab sums it up: the solution “would have to be relatively pure as other compounds also absorb at 280nm.”
What is settled and what is still an estimate
The physics is settled. Tryptophan and tyrosine side chains absorb near 280 nm, with a small contribution from cystine, and absorbance rises in a straight line with concentration across a working range. For a pure protein of known sequence, A280 is fast, needs no reagent, and leaves the sample intact.
The numbers built on top are averages. The Pace coefficients describe a typical folded protein, not yours. Pace’s group found that tryptophan absorbs less inside proteins than in any solvent they tested, and tyrosine falls between its values in denaturant and in propanol. Exactly how much each residue absorbs depends on its surroundings in the folded protein, and a sequence alone can’t capture that. The 0.57 ratio, the Layne correction and the 333 nm scattering fix are useful working tools, but none is a law.
Pace’s group concluded that “the best approach is to measure rather than predict” a protein’s coefficient, and the Basel guide recommends calibrating against amino acid analysis for the highest accuracy. For a protein a lab works with routinely, measuring its own coefficient once, by the Edelhoch method or by amino acid analysis, replaces an average-based estimate with a value specific to that protein.