Quantifying Protein Expression

Protein expression occurs in a variety of experimental, natural, and production environments, and many ways have been invented to describe “expression levels.” All involve some aspect of quantifying protein concentrations, but not all such measurements are done the same way. The most relevant category on which to base analysis is protein purity.

Most assays of pure proteins are based on standard analytical methods that including traditional (i.e., non-mass) spectroscopy, chromatography, and wet chemistry. Many proteins, particularly enzymes, may be quantified through functional assays that measure activity and even there the units are non-standard. Activity assays may be reported alone (as “units” instead of mass), or with mass calculations to provide specific activity.

A non-analytical scientist tends to take results from these assays at face value, and the numbers they generate as absolute. One reason so many protein analytical methods exist, however, is the diversity of sources, circumstances, and end-uses.

“All methods currently in use today have benefits and drawbacks,” says Tim Sontag, Ph.D., Principal Scientist at Proteintech, which specializes in custom protein production and analytics, “and all such methods are imperfect.”

Some of the most common protein quantitation methods take advantage of a “typical” protein’s absorbance in the ultraviolet absorbance range, specifically at 280 nm. The number of UV-absorbing aromatic (tyrosine, tryptophan, phenylalanine) amino acid residues contained in the protein determine the molecule’s extinction coefficient (molar absorption), which can then be directly correlated with concentration in mg/mL.

“This method is convenient due to its ease and accessibility, with no additional reagents required,” Sontag says. “Absorption measurements are only applicable to highly pure samples solubilized in a buffer that does not absorb at 280 nm.”

Another common protein assay is the Biuret method, which works through chelation of the target protein to copper and subsequent measurement of reduced copper. Readouts are compared to the concentration curve for a standard protein (e.g., bovine serum albumin). Sontag notes that this assay does not work in the presence of reducing agents.

The third most common method, employing protein-dye interactions, works by detecting color changes in the dye as a function of protein concentrations. This assay is inhibited by the presence of certain detergents that interfere with protein-dye interactions.

“The problem for us is the same protein sample can give widely varying results even when no inhibitory factors are present,” Sontag notes. For example, when measuring levels of the common cytokine IL-7, you may get a protein concentration of 1 mg/mL by absorbance at 280 nm, but the Biuret reading may only be 0.5 mg/mL. At the same time, for another cytokine, IL-8, the readouts from the two methods will be the same.

“So if a lab is looking for simplified protein quantitation, the solution is not to seek a perfect solution, but rather to understand the drawbacks of the method chosen and to be consistent in your choice of method. In the case of Proteintech, we package our purified proteins based on the protein quantities in milligrams. We measure our proteins by both Abs280 and by the BCA (bicinchoninic acid) method, a type of Biuret assay, and label the product based on the BCA results. This way, when a customer purchases a 1 mg from us today, or 5 years from now, the actual amount in the vial will remain the same.”

Complex samples, complex analysis

Note that these methods are only appropriate for measuring either total protein in a sample or purified proteins.

Mixtures of proteins, such as one encounters in cell culture-based or microbial biomanufacturing, require either antibody-based quantitation directed at the specific protein of interest, or mass spectral analysis. Both approaches carry the additional drawbacks of taking significant time, relying on highly specific antibody reagents, and sometimes requiring significant sample prep, according to Sontag.

“Sample prep for mass spec involves protein digestion to look at fragmentation patterns. The analysis of the results is more complex than typical protein quantitation methods. Antibody-based protein quantitation involves far less sample prep but results strongly depend on antibody specificity and the protein standard used.”

Which is why, although perhaps it is not always the preferred method for simple protein quantitation, mass spectrometry (MS) can help quantify expression levels of specific polypeptides swimming in a veritable protein soup. In addition, it can help quantify relative levels of isoforms and provide additional structural information.

Generally, protein quantitation via MS involves monitoring for ions that are specific to the protein (or peptide) of interest and comparing those responses against a standard curve generated with synthetic peptides.

To control for protein instability under biological conditions Michael Bacica, Sr. Scientist, Mass Spectrometry at the Pelican Expression Technology™ group of Ligand Pharmaceuticals, generates serum stability curves on therapeutic peptides to generate half-life values. “We generate a standard curve with the peptide itself, using ions specific for the peptide and two fragment ions from the peptide. We have obtained excellent results from samples where the total ion chromatogram (TIC) was unbelievably complex due to the serum’s normal protein content. For larger proteins, one would digest the sample with an enzyme to get to the peptides to monitor. As an example, for an antibody you would monitor peptides containing complementarity-determining regions, which are specific for that protein.”

Host cell proteins (HCPs) are also expressed by cells alongside product proteins and need to be tracked and quantified, which Ligand accomplishes through a semi-quantitative liquid-chromatography MS (LC-MS) method. Their technique, involving spiking with known quantities of standard proteins, tryptic digestion, and depletion of the drug substance, compares the HCP levels against 25, 100, or 200 ppm standard peptides.

Quantifying product proteins in complex purification feed streams is essential for developing downstream processes. Bacica uses LC or microfluidic separation techniques, followed by intact mass analysis, on proteins of interest at different concentrations, generating a curve based on the differing ion intensities. “From there we can go to an affinity capture method, like Protein A (for antibody scaffolds) or IMAC (for his-tagged proteins) and generate a value for capturable titer for proteins present in a crude lysate.”

Quantifying protein expression holds interest from many angles, from huge cell cultures to single cells, from high school experiments to drug manufacturing. Systematic, experimental, and human error all contribute to the uncertainty of the final numbers but the goal always is consistency so that numbers, while imprecise in an absolute way, are nevertheless comparable.