I agree with everyone above, but would like to clarify the relative contributions and "reality" due to error in measurement. The absorbance at 280 is not dependent on glycosylation, beyond a tiny change that could occur because of masking W, Y or C from solvent. In addition to the contributions of W, Y and C (which are not equivalent as can be seen here: ε = (nW×5500) + (nY×1490) + (nC×125); the contribution from W is ~50X more than an equivalent amount of C for example), there is a tiny contribution from other things such as F and the peptide bond too. Regardless, most people use the calculated extinction coefficient to estimate the protein concentration, because the error associated with the differences due to conformation, etc is as large as the error in any other common way of measuring protein concentration such as a Bradford assay.

As far as I know, the UV absorbance of saccharides (and that means also of glycans) is negligible. Regarding conformation - the UV absorption should not be influenced by conformation. So, by measuring UV absorbance you measure the proteins, or the protein part of glycoproteins ...

As Sterba said UV the absorbance of saccharides (at 280 at least) is negligible. Other topic is the protein conformation. As protein conformation determine the amount of Trp, Tyr and Cys residues buried or exposed it will influence the absorbance. If the protein denaturalize and unfold you can expect the same amount of protein to absorb more at 280, and you can not longer relay in its molar extinction coefficient for quantification.

Saccharides should not absorb in UV-VIS whatsoever.

There is difference with the conformation, because electron environment affect the absorbance.

(similarly changes absorbance of DNA after denaturation, because the nucleotides are in different environment)

Hydrocarbons do not show conjugate double bonds making their absorption fall in the UV-Vis range. Though they may slightly change dielectrics at their surroundings, their effect on the absorption of other residues must be negligible and very dependent on the relative orientation of the absorbent and the hydrocarbon rings. Conformation affects absorption (changing the environment of aromatic rings from polar to non-polar shifts the absorption towards blue),fluorescence (motulating energy transfer and quenching) and dychroism (anysotropic environment near amides and aromatic rings).

I agree with everyone above, but would like to clarify the relative contributions and "reality" due to error in measurement. The absorbance at 280 is not dependent on glycosylation, beyond a tiny change that could occur because of masking W, Y or C from solvent. In addition to the contributions of W, Y and C (which are not equivalent as can be seen here: ε = (nW×5500) + (nY×1490) + (nC×125); the contribution from W is ~50X more than an equivalent amount of C for example), there is a tiny contribution from other things such as F and the peptide bond too. Regardless, most people use the calculated extinction coefficient to estimate the protein concentration, because the error associated with the differences due to conformation, etc is as large as the error in any other common way of measuring protein concentration such as a Bradford assay.

I am not sure of the effect of glycosylation but non standard proteins can be measured more accurately by different wavelengths.  Rather than benzene rings of Try etc. double bonds can be measured using 230 -260, I think the formula is (183x A230)- (25.6xA260) so these wavelengths will neatly overlap with the 280/260 assay for BSA but is more sensitive and will be more accurate with the non-standard proteins low in aromatic rings .  This and other assays using wavelengths at A205 for example may provide a more user friendly assay with validation.