As the subject heading says, we have a 'protein' present in a fraction from a chromatography run, but no peak at 214nm for the same fraction (we do get peaks elsewhere and can identify these). We have done a number of tests which have more or less confirmed that there is protein present in the fraction, as well as other very big molecules/aggregates, such as DNA (sorry, I can't go into much more detail on this). These tests strongly suggest that there should be enough material present to show up at 214nm, but we get nothing at all.
There is nothing in the buffers we are running that is likely to interfere at this wavelength, and all other peaks are as expected.
Is it possible to mask a 214nm signal? We know that it is possible to miss a protein at 280nm if the aromatic amino acids that absorb at this wavelength are on the inside of the protein and/or hidden within micelles or other components. Does the same apply to the 214nm wavelength used to detect the peptide bond? Is it possible that a protein complexed with DNA, or internalised within lipid micelles or other large molecules/aggregates can be invisible at various UV wavelengths? Is there a technique that could be used to overcome this problem?

One usually consideres the influence of the medium on absorbance as relatively (for instance in comparison to fluorescence) small. The shifts in wavelengths (lamda max) due to changes in polarity, etc., are well documented. These cause, of course, a change of absorbance at some wavelengths (usually not much at the shifted lambda max). Therefore, one would expect that unless your protein concentration is at the detection limit you should see it if it emerges from the column. If you can prove a considerable "masking" it would shurely be a highly interesting result.
This should also hold for 280nm. Do you have unequivocal evidence for an absorption "masking" at 280? Are you sure that this is not a scattering problem (so much scattering that the absorbance is negligible)? It would certainly be nice to have a spectroscopist´s opinion on this.

It seems to me that I have seen something on studying macromolecule conformation with absorbance (fluorescence is extensively used in this), but the memory is very nebulous as I assumed that this would be a rare special case. If I am wrong it would be vey important to find out.

I would check if you haven't missed the target compound in the fraction collection. There is a delay between the detector and the point where the fraction is collected due to tubing etc. Since you got a lot of gunk in the fraction, you can't tell if you collected the right fraction or a fraction before the peak.
Other wild things are possible too, but I would start first making sure that I did not make a mistake.

To anon: The 'peakless' fraction occurs in several chromatography systems, FPLC prepararative & analytical & HPLC columns, the fraction in question is well away from any other peaks and delays have been taken into account either by the system or the operator.
To HWM: It is possible that the protein is at the detection limit, but the tests used subsequently on the fraction would suggest that we should see it. As for the light scattering issue, I think this is more likely to be the cause but we haven't looked at it yet (mainly because we don't have multiple wavelength detectors). We may well carry out some runs at different wavelengths to see if there is an effect. We often use 320nm, 380nm or 580nm for light scattering (and other non-UV wavelengths) - are these suitable? What should we look for?
As for the evidence for masking at 280nm, I'm afraid it's nothing more than anecdotal evidence from our lab (protein disappearing in the presence of detergents above their CMC - which would suggest light scattering). My understanding is that in some cases you will get different results when determining the extinction coefficient of a native protein compared to the denatured protein, the theory being that if there are internal aromatic amino acids these are not as visible as those on the surface of the protein - does this hold?
Clearly, spectroscopy is not my field.

What UV detector are you using for the detection? It appears that it is not a standard multi-wavelength UV. If it is not, the question is if you see anything at 214nm. Maybe your detector is blind. If this is the case, the next question is the mobile phase that you are using. Some mobile phase combinations do not let any light go through at 214.
Is this a possible path?

David,
To check on scattering you would use a wavelength where your substance does not absorb, but remember that scattering depends on lumbda^-4, so shorter wavelength scatter more stongly. Actually, to obliterate a peak via scattering you would have to have something that scatters come out of the column more or less continuously. (In the words of Anon., the detector would have to be blinded)
On the use of absorbance for structural info: I found the material that lingered in my thoughts: D. Freifelder, Physical Biochemisty, Freeman, NY, 1982, p. 500. Very brusquely: Polarity of the surroundings of a chromophore (solvent or parts of the macro molecule) can shift the absorption, but not extremely, pH change of the chromophore can cause large shifts. Still, I would not expect the disappearance of an absorption band unless one is at the low end of detection to start with.

Many thanks for the responses.
I have a feeling that we are at the detection limits for our protein of interest. I mentioned that our other tests have suggested that we should be able to see the peak, but I think we were perhaps over-interpreting them. We probably have one assay system (chromatography) which is not quite sensitive enough and another which is 'oversensitive'. The same response in the other assay system for other proteins has given us identifiable peaks in various chromatography systems, and we essentially applied the same rules to this fraction. It's all a bit unfortunate because we wanted to have a single assay for this 'protein' but will probably need to combine it with the sensitive assay (the chromatography step is needed anyway as a clean-up step for the sensitive assay).
As for all this talk on lambda max, etc, - it has made me realise just how little I know on the subject, so I'll be hitting the text books again...